]> code.ossystems Code Review - openembedded-core.git/log
openembedded-core.git
14 years agocmake.bbclass: sync from openembedded
Qing He [Fri, 20 Aug 2010 02:57:48 +0000 (10:57 +0800)]
cmake.bbclass: sync from openembedded

this fixes:
  - toolchain detection error (e.g. ar and ranlib)
  - cmake modules search path

Signed-off-by: Qing He <qing.he@intel.com>
14 years agogawk: add 3.1.5 (GPLv2 version)
Yu Ke [Wed, 25 Aug 2010 09:15:48 +0000 (17:15 +0800)]
gawk: add 3.1.5 (GPLv2 version)

this commit add GNU Awk 3.1.5, the most recent GPLv2 version.
the recipe is from OE, and upgrade to 3.1.5

Signed-off-by: Yu Ke <ke.yu@intel.com>
14 years agocpio: new recipe v2.8
Scott Garman [Tue, 17 Aug 2010 01:55:33 +0000 (18:55 -0700)]
cpio: new recipe v2.8

This is the last GPLv2 version of cpio. The recipe was derived from
OE's v2.5 one.

Signed-off-by: Scott Garman <scott.a.garman@intel.com>
14 years agoat.inc: remove unneeded LIBS setting
Scott Garman [Mon, 16 Aug 2010 22:00:38 +0000 (15:00 -0700)]
at.inc: remove unneeded LIBS setting

Signed-off-by: Scott Garman <scott.a.garman@intel.com>
14 years agotar_1.17: eliminate a patch by setting CFLAGS instead
Scott Garman [Mon, 16 Aug 2010 21:45:29 +0000 (14:45 -0700)]
tar_1.17: eliminate a patch by setting CFLAGS instead

Signed-off-by: Scott Garman <scott.a.garman@intel.com>
14 years agofile: reinstate 5.04 upgrades with modifications
Qing He [Tue, 17 Aug 2010 02:58:15 +0000 (10:58 +0800)]
file: reinstate 5.04 upgrades with modifications

reinstate the patch since rpm recipe has been upgraded

> commit 21d586b47c8c5338913c5e8ca2ee604409cc57a0
> Author: Qing He <qing.he@intel.com>
> Date:   Tue Jul 6 13:19:10 2010 +0800
>
>     file: upgrade to version 5.04
>
>     from 4.18
>
>     changes:
>       - remove native-fix.patch and sed line in do_configure_prepend,
>         AM_CONDITIONAL(IS_CROSS_COMPILE, ...) works well
>       - fix the bug of not packaging ${datadir}/misc/magic.mgc,
>         without this file, this utility isn't very useful
>       - add zlib as explicit build dependency
>
>     Signed-off-by: Qing He <qing.he@intel.com>

new changes:
  - string/B is not supported in file 5.04, add an alias to string/W
  - '>=' and '<=' was not supported, add an extension for existing
    usage

Signed-off-by: Qing He <qing.he@intel.com>
14 years agolibuser: add 0.56.16 as new recipe
Zhai Edwin [Wed, 25 Aug 2010 08:05:35 +0000 (16:05 +0800)]
libuser: add 0.56.16 as new recipe

Signed-off-by: Zhai Edwin <edwin.zhai@intel.com>
14 years agobase.bbclass: Add xz-native as depends when finding *.xz in SRCURI
Zhai Edwin [Wed, 25 Aug 2010 08:05:13 +0000 (16:05 +0800)]
base.bbclass: Add xz-native as depends when finding *.xz in SRCURI

Just adding xz-native as DEPENDS in one bb file could not make its unpack run
after xz-native build done.

Signed-off-by: Zhai Edwin <edwin.zhai@intel.com>
14 years agoxz: add 4.999.9beta as new recipe
Zhai Edwin [Sun, 22 Aug 2010 03:00:42 +0000 (11:00 +0800)]
xz: add 4.999.9beta as new recipe

Requies xz-native to unpack the tarball of libuser

Signed-off-by: Zhai Edwin <edwin.zhai@intel.com>
14 years agobash (GPLv2): remove metadata sharing from GPLv3 recipe
Kevin Tian [Tue, 24 Aug 2010 07:34:35 +0000 (15:34 +0800)]
bash (GPLv2): remove metadata sharing from GPLv3 recipe

Signed-off-by: Kevin Tian <kevin.tian@intel.com>
14 years agofindutils (GPLv2): remove metadata sharing from GPLv3 recipe
Kevin Tian [Tue, 24 Aug 2010 07:17:20 +0000 (15:17 +0800)]
findutils (GPLv2): remove metadata sharing from GPLv3 recipe

to avoid inadvertent contamination in the future, GPLv2 recipe is
fully split its v3 buddy. This way we can enforce strong isolation
for GPLv2 code.

also remove one unnecessary patch

Signed-off-by: Kevin Tian <kevin.tian@intel.com>
14 years agorsync: split GPLv2 version from GPLv3 by fixing the GPLv2 recipe
Dexuan Cui [Tue, 24 Aug 2010 03:34:29 +0000 (11:34 +0800)]
rsync: split GPLv2 version from GPLv3 by fixing the GPLv2 recipe

remove the "require rsync.inc", and copy the .inc file into the GPLv2 recipe,
and adjust it properly;

mkdir rsync-2.6.9 and copy files/rsyncd.conf into rsync-2.6.9/rsyncd.conf.
BTW: the .conf file originated from OE in 2006:

commit c2a4b021eb04924f8c8b0e0af1fac337a04cbcae
Author: Raymond Danks <raymond@edanks.com>
Date:   Tue Apr 18 01:43:08 2006 +0000

    rsync: Add rsyncd.conf

http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?id=c2a4b021eb04924f8c8b0e0af1fac337a04cbcae

Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
14 years agowhich: add GPLv2 version 2.18
Dexuan Cui [Fri, 20 Aug 2010 12:40:23 +0000 (20:40 +0800)]
which: add GPLv2 version 2.18

Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
14 years agocoreutils: upgrade to latest GPLv2 version (6.9)
Kevin Tian [Wed, 18 Aug 2010 04:55:21 +0000 (12:55 +0800)]
coreutils: upgrade to latest GPLv2 version (6.9)

Before adding latest coreutils GPLv3 version, first we upgrade existing
recipe to latest GPLv2 version here.

To ensure a clean process, all bits are coded from scratch including
patches and recipe, except that in recipe lines about update-alternative
are reused from previous 5.3.0. All new patches are licensed under
GPLv2.

Tested simple features with most coreutils commands

Signed-off-by: Kevin Tian <kevin.tian@intel.com>
14 years agotest: Add scenario file for each target and support single case running
Jiajun Xu [Wed, 18 Aug 2010 14:02:08 +0000 (22:02 +0800)]
test: Add scenario file for each target and support single case running

Different test cases are needed for different targets. A folder "scenario"
is created under scripts/qemuimage-tests to hold scenario files for different
targets.

Single case running is supported now. User can run single case together with
a whole test suite by setting variable TEST_SCEN in local.conf.

By default test cases in sanity suite will be ran. If you want to run other
test suite or specific test case(e.g. bat or boot test case under sanity suite),
list them like following.
TEST_SCEN = "sanity bat sanity:boot"

Signed-off-by Jiajun Xu <jiajun.xu@intel.com>

14 years agotask-poky-basic: added new packages to basic task for LSB
Saul Wold [Mon, 16 Aug 2010 20:44:21 +0000 (13:44 -0700)]
task-poky-basic: added new packages to basic task for LSB

further tweaked by Kevin Tian, with simple boot test. 'at' is disabled
as it depends on mail of build system which needs to be fixed

Signed-off-by: Saul Wold <Saul.Wold@intel.com>
Signed-off-by: Kevin Tian <kevin.tian@intel.com>
14 years agoftrace: add trace-cmd for ftrace control and config
Bruce Ashfield [Fri, 20 Aug 2010 16:24:43 +0000 (12:24 -0400)]
ftrace: add trace-cmd for ftrace control and config

Rather than manually configuring ftrace, trace-cmd was created
to provide a more user friendly interface to the tracer.

Some examples are:

 # ./trace-cmd record -e all ls /bin

Then to see the information:

 # ./trace-cmd report

To see just a subsystem:

 # ./trace-cmd record -e irq ls /bin

To list the available trace events:

 # ./trace-cmd list -e

You can also trace a single trace point with one of two methods:

 # ./trace-cmd record -e sched:sched_wakeup ls /bin

Or simply:

 # ./trace-cmd record -e sched_wakeup ls /bin

This tool also works with tracers (plugins) as well as events

 # ./trace-cmd list -p

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
14 years agolinux-wrs: update to 2.6.34.5
Bruce Ashfield [Thu, 26 Aug 2010 18:09:10 +0000 (14:09 -0400)]
linux-wrs: update to 2.6.34.5

edd8801 Linux 2.6.34.5
4070f43 vmware: fix build error in vmware.c
57cbde5 mm: fix up some user-visible effects of the stack guard page
4def7ce mm: fix page table unmap for stack guard page properly

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
14 years agosstate/utility-tasks: Ensure do_clean functions correctly and removes shared state
Richard Purdie [Fri, 27 Aug 2010 11:33:58 +0000 (12:33 +0100)]
sstate/utility-tasks: Ensure do_clean functions correctly and removes shared state

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agosstate: Run sstate_cleanall() as part of do_setscene()
Richard Purdie [Fri, 27 Aug 2010 11:29:01 +0000 (12:29 +0100)]
sstate: Run sstate_cleanall() as part of do_setscene()

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agoqemu: On some older machines SDL is available
Mark Hatle [Thu, 26 Aug 2010 18:49:37 +0000 (11:49 -0700)]
qemu: On some older machines SDL is available

On some older machines SDL is available, even if a pkgconfig
file for SDL is not there.  (See RHEL5.1)

Extend the check, to see if the primary SDL header is there, if
the pkgconfig check fails.

Signed-off-by: Mark Hatle <mhatle@windriver.com>
14 years agopseudo: Update to the latest git version of pseudo
Mark Hatle [Thu, 26 Aug 2010 18:45:55 +0000 (11:45 -0700)]
pseudo: Update to the latest git version of pseudo

Latest git version includes:

2010-08-26:
* (seebs) make offsets.c slightly less useless
* (seebs) don't overwrite LD_LIBRARY_PATH values that include us

2010-08-25:
* (seebs) fix the signal mask restore

2010-08-24:
* (seebs) try to restore signal mask before calling exec()
* (seebs) move errno restoration after code which could set errno

2010-08-19:
* (seebs) handle insane edge case involving regcomp/regexec

2010-08-17:
* (seebs) create speculative-deletion logic
* (seebs) remove crackpot theories about cross-device renames

2010-08-16:
* (rp) Fix ld_preload/ld_library_path mixup.
* (seebs) Handle failed allocations.
* (seebs) Fix logic for dropping empty LD_PRELOAD.

Signed-off-by: Mark Hatle <mhatle@windriver.com>
14 years agometa-toolchain: fix for recent opkg change
Joshua Lock [Thu, 26 Aug 2010 17:01:49 +0000 (18:01 +0100)]
meta-toolchain: fix for recent opkg change

Signed-off-by: Joshua Lock <josh@linux.intel.com>
14 years agometa-toolchain: update to reflect opkg state directory move
Joshua Lock [Thu, 26 Aug 2010 07:59:20 +0000 (08:59 +0100)]
meta-toolchain: update to reflect opkg state directory move

Signed-off-by: Joshua Lock <josh@linux.intel.com>
14 years agoopkg: finalize the change to var directory
Kevin Tian [Thu, 26 Aug 2010 04:39:07 +0000 (12:39 +0800)]
opkg: finalize the change to var directory

commit 1d0757f16beb31551733d9d755d72337ccda9642 changes opkg run state from
/usr/lib/opkg to /var/lib/opkg, which however is incomplete and still many
important information is kept under old directory including postinst methods.
This makes latest boot into a mess.

So finalize this movement to /var here. Fix [BUGID #229]

Signed-off-by: Kevin Tian <kevin.tian@intel.com>
14 years agosstate: fix architecture of sstate archives
Joshua Lock [Wed, 25 Aug 2010 14:00:02 +0000 (15:00 +0100)]
sstate: fix architecture of sstate archives

Using TARGET_ARCH as the pacakge architecture for sstate archives lead to some
issues when using two machines with different BASE_PACKAGE_ARCH's but the same
TARGET_ARCH such as netbook and qemux86 machines.

Probable fix for [BUGID #223]

Signed-off-by: Joshua Lock <josh@linux.intel.com>
14 years agoxf86-video-intel: fix COMPATIBLE_HOST entry, this was preventing building
Joshua Lock [Wed, 25 Aug 2010 10:37:29 +0000 (11:37 +0100)]
xf86-video-intel: fix COMPATIBLE_HOST entry, this was preventing building

Signed-off-by: Joshua Lock <josh@linux.intel.com>
14 years agopoky-qemu: check for existence of required libraries
Joshua Lock [Tue, 24 Aug 2010 15:16:47 +0000 (16:16 +0100)]
poky-qemu: check for existence of required libraries

We need some libraries installed to run the built qemu, check for their
existence before trying to run the program. Fixes [BUGID #201]

Signed-off-by: Joshua Lock <josh@linux.intel.com>
14 years agoqemu.inc: ensure helper programs are built when building for qemu machines
Joshua Lock [Tue, 24 Aug 2010 13:47:10 +0000 (14:47 +0100)]
qemu.inc: ensure helper programs are built when building for qemu machines

QEMU machine images are useless without qemu-native and qemu-helper-native

Signed-off-by: Joshua Lock <josh@linux.intel.com>
14 years agoopkg: use var for state files, not libdir
Joshua Lock [Tue, 24 Aug 2010 10:32:01 +0000 (11:32 +0100)]
opkg: use var for state files, not libdir

Fixes [BUGID #221]

Signed-off-by: Joshua Lock <josh@linux.intel.com>
14 years agobitbake/runqueue.py: Ensure rqexe always exists and that empty task lists cause a...
Richard Purdie [Tue, 24 Aug 2010 23:58:23 +0000 (00:58 +0100)]
bitbake/runqueue.py: Ensure rqexe always exists and that empty task lists cause a graceful exit

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agopackage_rpm: Temporary workaround for perfile dependencies
Mark Hatle [Mon, 23 Aug 2010 15:12:43 +0000 (08:12 -0700)]
package_rpm: Temporary workaround for perfile dependencies

Until the pseudo code is fully integrated, we need to roll up the
per file dependencies into package dependencies for rpm.

Signed-off-by: Mark Hatle <mhatle@windriver.com>
14 years agorootfs_rpm: Enable multi dep solverdbs
Mark Hatle [Sun, 22 Aug 2010 05:44:21 +0000 (22:44 -0700)]
rootfs_rpm: Enable multi dep solverdbs

Update package_rpm.bbclass to generate dep solver databases for each
package arch.  Following the example of the deb and ipk integration

Revise the rootfs_rpm to solve the installation based on the multiple
dep solvers.

Note, recommends, locale and attemptonly pckages are still to be
implemented.

Signed-off-by: Mark Hatle <mhatle@windriver.com>
14 years agopackage_rpm: Rewrite the way the spec files are generated
Mark Hatle [Fri, 20 Aug 2010 16:52:39 +0000 (09:52 -0700)]
package_rpm: Rewrite the way the spec files are generated

Use a single spec file to generate all of the split packages.  This allows
us to ensure the RPM package source package meta data is correct, and also
speeds up the package generation process.

Signed-off-by: Mark Hatle <mhatle@windriver.com>
14 years agopackage.bbclass: Fix per-file dependency generation
Mark Hatle [Mon, 23 Aug 2010 16:12:54 +0000 (09:12 -0700)]
package.bbclass: Fix per-file dependency generation

The overall file list was being constructed with the wrong variable, it
should have had FLIST in the name.

Also it was possible to construct the system with some illegal variable
names.  Names that would have include "[]".  So translate these away...

Finally fix an issues where a filename with an _ could cause a package
variable conflict.  Again translate this away..

Signed-off-by: Mark Hatle <mhatle@windriver.com>
14 years agorpm_5.1.10.bb: Update macros
Mark Hatle [Sun, 22 Aug 2010 02:40:43 +0000 (19:40 -0700)]
rpm_5.1.10.bb: Update macros

Fix some additional macro definitions to enable rpmdeps usage

Signed-off-by: Mark Hatle <mhatle@windriver.com>
14 years agorpm: use rpm-5.1.10 from SNAPSHOT URI.
Jeff Johnson [Sat, 21 Aug 2010 03:09:44 +0000 (23:09 -0400)]
rpm: use rpm-5.1.10 from SNAPSHOT URI.

rpm-5.1.10 snapshot adds support for multiple aid databases

revert slashDepth patch, now included in 5.1.10

Signed-off-by: Jeff Johnson <n3npq@mac.com>
Signed-off-by: Mark Hatle <mhatle@windriver.com>
14 years agoEnsure all packages which should have x86_64 in COMPATIBLE_HOST
Joshua Lock [Wed, 18 Aug 2010 17:54:14 +0000 (18:54 +0100)]
Ensure all packages which should have x86_64 in COMPATIBLE_HOST

Signed-off-by: Joshua Lock <josh@linux.intel.com>
14 years agopackage.bbclass: Fix shlibs race
Richard Purdie [Mon, 23 Aug 2010 14:41:21 +0000 (15:41 +0100)]
package.bbclass: Fix shlibs race

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agobitbake.conf: Add DESCRIPTION for -dev, -dbg and -doc packages
Richard Purdie [Mon, 23 Aug 2010 14:27:18 +0000 (15:27 +0100)]
bitbake.conf: Add DESCRIPTION for -dev, -dbg and -doc packages

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agoinsane.bbclass: Update to match the packaging classes when handling OVERRIDES
Richard Purdie [Mon, 23 Aug 2010 14:26:38 +0000 (15:26 +0100)]
insane.bbclass: Update to match the packaging classes when handling OVERRIDES

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agopackage.bbclass: Fix do_shlibs
Mark Hatle [Sat, 21 Aug 2010 19:21:02 +0000 (12:21 -0700)]
package.bbclass: Fix do_shlibs

The current packages shlibs processing is written to a local temporary file.
This file is not consulted when evaluting the provider of the dependency,
instead we need to cache the local values into the dictionary while we write
the temp file.

Signed-off-by: Mark Hatle <mhatle@windriver.com>
14 years agodevshell.bbclass: export the FAKEROOTENV so it can be accessed within the devshell...
Richard Purdie [Mon, 23 Aug 2010 12:04:03 +0000 (13:04 +0100)]
devshell.bbclass: export the FAKEROOTENV so it can be accessed within the devshell task

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agoutils.py: Fix bb.copyfile to change the permissions of the file back correctly
Richard Purdie [Mon, 23 Aug 2010 12:01:05 +0000 (13:01 +0100)]
utils.py: Fix bb.copyfile to change the permissions of the file back correctly

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agopackage.bbclass: If files are unreadable, add read access for stripping
Richard Purdie [Sat, 21 Aug 2010 23:28:28 +0000 (00:28 +0100)]
package.bbclass: If files are unreadable, add read access for stripping

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agolib/oe/path: Use bb.utils.copyfile as shutils can't cope with copying unreadable...
Richard Purdie [Sat, 21 Aug 2010 23:27:55 +0000 (00:27 +0100)]
lib/oe/path: Use bb.utils.copyfile as shutils can't cope with copying unreadable files

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agoutils.py: Add a new extend_deps() to easily merge two dependency lists
Mark Hatle [Fri, 20 Aug 2010 16:27:22 +0000 (09:27 -0700)]
utils.py: Add a new extend_deps() to easily merge two dependency lists

Add a new extend_deps function to more easily merge two dependency lists.
This avoids adding duplicates, unless the value of the dependency is
different.

Signed-off-by: Mark Hatle <mhatle@windriver.com>
14 years agopackage.bbclass: Add perfile dependency identification
Mark Hatle [Fri, 20 Aug 2010 15:54:53 +0000 (08:54 -0700)]
package.bbclass: Add perfile dependency identification

Using the rpmdeps commands, we enable the discovery of per-file dependencies.
These dependencies are generated for all packages, and may or may not be used
by any given packaging mechanism.  Currently RPROVIDES and RDEPENDS are
generated in this process.

Signed-off-by: Mark Hatle <mhatle@windriver.com>
14 years agobitbake.conf: Add shared/pkgconfig files to standard -dev glob
Mark Hatle [Fri, 20 Aug 2010 15:38:04 +0000 (08:38 -0700)]
bitbake.conf: Add shared/pkgconfig files to standard -dev glob

Add the pkgconfig files that may appear into the shared directory into
the -dev globbing.

Also change the udev integration to remove the manual instance of the
shared directory .pc file.

Signed-off-by: Mark Hatle <mhatle@windriver.com>
14 years agorpm: Update the rpm integration to support per-file dependencies
Mark Hatle [Fri, 20 Aug 2010 15:32:08 +0000 (08:32 -0700)]
rpm: Update the rpm integration to support per-file dependencies

Update the RPM package integration to support per-file dependencies

This adds additional configuration options to RPM, as well as provides
a helper script "perfile_rpmdeps.sh" that the build system can use to
gather the dependency information.

Signed-off-by: Mark Hatle <mhatle@windriver.com>
14 years agoglibc: bring back needed patches from older recipes
Nitin A Kamble [Fri, 20 Aug 2010 19:20:52 +0000 (12:20 -0700)]
glibc: bring back needed patches from older recipes

while removing old glibc recipes some needed patches from
older recipes got wiped out. Bringing them back.

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
14 years agobase.bbclass: Initial work to add GPLv2 Only Build with Whitelist for GPLv3 Native...
Saul Wold [Mon, 16 Aug 2010 21:48:04 +0000 (14:48 -0700)]
base.bbclass: Initial work to add GPLv2 Only Build with Whitelist for GPLv3 Native tools

[tweaks from Richard Purdie]

Signed-off-by: Saul Wold <Saul.Wold@intel.com>
Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agolinux-wrs: allow packed refs to be present
Bruce Ashfield [Thu, 19 Aug 2010 20:37:03 +0000 (16:37 -0400)]
linux-wrs: allow packed refs to be present

The code which creates a working tree from the source
Wind River Linux git repository manipulates the refs
to make non-tracking, local branches. packed-refs remove
the files from refs, making this code break and the builds
fail.

To fix this, if we detect the packed refs we simply create
the local refs from the remote branches.

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
14 years agokernel: update branch revisions for 2.6.34.4
Bruce Ashfield [Thu, 19 Aug 2010 20:29:36 +0000 (16:29 -0400)]
kernel: update branch revisions for 2.6.34.4

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
14 years agoguilt: import guilt for kernel patch management
Bruce Ashfield [Tue, 10 Aug 2010 19:01:37 +0000 (15:01 -0400)]
guilt: import guilt for kernel patch management

The Wind River kernel is "patched" via guilt to provide
both git integration and quilt like patch management
(if that is of interest).

This is a modified 0.33 guilt with some changes to
streamline interactions with the way that the Wind
River kernel is constructed. That being said, the
common semantics of guilt are not changed, and it
can be used for other purposes.

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
14 years agodistro-tracking: add info for watchdog
Dexuan Cui [Fri, 20 Aug 2010 08:11:40 +0000 (16:11 +0800)]
distro-tracking: add info for watchdog

Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
14 years agowatchdog: add the latest version 5.9
Dexuan Cui [Fri, 20 Aug 2010 07:54:59 +0000 (15:54 +0800)]
watchdog: add the latest version 5.9

Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
14 years agopoky-qemu-internal: fix the echo of KERNCMDLINE
Dexuan Cui [Thu, 19 Aug 2010 07:01:29 +0000 (15:01 +0800)]
poky-qemu-internal: fix the echo of KERNCMDLINE

Without the fix, when the commandline is printed with the "echo" command,
it loses the quotes and this confuses people.

Thanks Richard for pointing this out.

Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
14 years agodistro tracking: Add info for pax
Zhai Edwin [Fri, 20 Aug 2010 07:22:15 +0000 (15:22 +0800)]
distro tracking: Add info for pax

Signed-off-by: Zhai Edwin <edwin.zhai@intel.com>
14 years agodistro tracking: Updated version info for owl-video, matchbox-desktop, webkit-gtk...
Zhai Edwin [Fri, 20 Aug 2010 07:22:15 +0000 (15:22 +0800)]
distro tracking: Updated version info for owl-video, matchbox-desktop, webkit-gtk...

Signed-off-by: Zhai Edwin <edwin.zhai@intel.com>
14 years agopoky-qemu-internal: fix broken call to poky-qemu-ifup
Scott Garman [Wed, 18 Aug 2010 18:10:14 +0000 (11:10 -0700)]
poky-qemu-internal: fix broken call to poky-qemu-ifup

The poky-qemu-ifup script now requires a path to the native
sysroot as an argument. This fixes a case where the argument
was missing.

Also, set up NATIVE_SYSROOT_DIR when running runqemu.

Signed-off-by: Scott Garman <scott.a.garman@intel.com>
14 years agopython-pycairo: upgrade to 1.8.10
Yu Ke [Mon, 16 Aug 2010 07:08:50 +0000 (15:08 +0800)]
python-pycairo: upgrade to 1.8.10

Signed-off-by: Yu Ke <ke.yu@intel.com>
14 years agoiputils: add iputils package
Yu Ke [Thu, 29 Jul 2010 07:43:22 +0000 (15:43 +0800)]
iputils: add iputils package

iputils are utilities for the IP protocol, including traceroute6, tracepath, tracepath6, ping, ping6 and arping.

this commit add iputils 20100418. code is borrowed from open embedded and debian, with following changes:
- upgrade from s20071127 to s20100418
- update the do_compile (add VPATH) to make s20100418 pass
- disable man since it depends docbook while poky don't have

Signed-off-by: Yu Ke <ke.yu@intel.com>
14 years agorunqemu-nfs: boot QEMU using nfsroot and local unfs export
Scott Garman [Fri, 13 Aug 2010 18:35:26 +0000 (11:35 -0700)]
runqemu-nfs: boot QEMU using nfsroot and local unfs export

This script automates the booting of QEMU using an nfsroot exported
by our userspace NFS tools. The rootfs should be created using
poky-extract-sdk.

Signed-off-by: Scott Garman <scott.a.garman@intel.com>
14 years agopoky-export-rootfs: script for exporting userspace nfsroots
Scott Garman [Fri, 13 Aug 2010 18:26:49 +0000 (11:26 -0700)]
poky-export-rootfs: script for exporting userspace nfsroots

This script automates the exporting of a root filesystem (created
with the poky-extract-sdk utility) using pseudo and the native
userspace NFS server. That filesystem can then be booted using
nfsroot with either QEMU or the target hardware using one of our
kernels.

Signed-off-by: Scott Garman <scott.a.garman@intel.com>
14 years agopoky-extract-sdk: script for extracting sdk images using pseudo
Scott Garman [Fri, 13 Aug 2010 18:12:33 +0000 (11:12 -0700)]
poky-extract-sdk: script for extracting sdk images using pseudo

This script automates the creation of a rootfs area using pseudo so
it can be used by a QEMU nfsroot boot.

Signed-off-by: Scott Garman <scott.a.garman@intel.com>
14 years agopoky-qemu-internal: fixes for userspace NFS booting
Scott Garman [Fri, 13 Aug 2010 18:18:05 +0000 (11:18 -0700)]
poky-qemu-internal: fixes for userspace NFS booting

Including a number of cases where the script could exit before
releasing a tun/tap lockfile.

Signed-off-by: Scott Garman <scott.a.garman@intel.com>
14 years agopoky-qemu-internal: Support use of a preconfigured tap device
Jeff Dike [Thu, 5 Aug 2010 19:13:45 +0000 (15:13 -0400)]
poky-qemu-internal: Support use of a preconfigured tap device

This patch makes poky-qemu-internal check for the existence of an
available preconfigured tap device before running poky-qemu-ifup to
make a new one.

Locking is handled with a lockfile in /tmp/qemu-tap-locks/.  This uses
the lockfile utility, so that needs to be present on the host.  On
exit, this script removes the lock file so that the tap device may be
reused.

Signed-off-by: Jeff Dike <jdike@linux.intel.com>
14 years agopoky-qemu-internal: Support NFS boots better
Jeff Dike [Thu, 5 Aug 2010 18:50:14 +0000 (14:50 -0400)]
poky-qemu-internal: Support NFS boots better

The image specification can now be in the form nfs-server:directory.
This makes it possible to nfs-boot from servers other than the host.
poky-qemu-internal will properly construct the kernel command line
given such a specification.

Signed-off-by: Jeff Dike <jdike@linux.intel.com>
14 years agopoky-qemu-[ifup|ifdown]: Use native tunctl
Jeff Dike [Thu, 5 Aug 2010 16:31:01 +0000 (12:31 -0400)]
poky-qemu-[ifup|ifdown]: Use native tunctl

poky-qemu-ifup can run standalone by root in order to configure a bank
of tap devices for later qemu use.

These devices will, if possible, be owned by a specified group to
which qemu users must belong.

If the kernel is too old to support TUNSETGROUP, then it falls back to
setting the tap device to be owned by a particular user, and that user
will be the only one allowed to use it.

Also overall usability improvements to the scripts, usage() help, etc.

Signed-off-by: Jeff Dike <jdike@linux.intel.com>
Signed-off-by: Scott Garman <scott.a.garman@intel.com>
14 years agopoky-find-native-sysroot: helper script for locating the native sysroot path
Scott Garman [Fri, 13 Aug 2010 17:50:37 +0000 (10:50 -0700)]
poky-find-native-sysroot: helper script for locating the native sysroot path

Various poky scripts make use of binaries from the native sysroot.
This helper script can be used to reduce code duplication, and sets
up some environment variables you can use to identify and obtain
the correct filesystem path to the native sysroot.

It works for both in-tree Poky setups as well as toolchain
installations.

Signed-off-by: Scott Garman <scott.a.garman@intel.com>
14 years agotunctl: Added an tunctl which supports TUNSETGROUP
Jeff Dike [Thu, 5 Aug 2010 15:29:35 +0000 (11:29 -0400)]
tunctl: Added an tunctl which supports TUNSETGROUP

TUNSETGROUP is needed in order to preconfigure a set of tap devices
that can be used by non-root users.  The requirement is that the qemu
users be members of whatever group the tap devices are assigned to.

Include tunctl in the qemu-helper package, and add a -native version.

Signed-off-by: Jeff Dike <jdike@linux.intel.com>
Signed-off-by: Scott Garman <scott.a.garman@intel.com>
14 years agopoky-qemu-ifup: allow multiple tap devices
Jeff Dike [Wed, 4 Aug 2010 21:19:59 +0000 (17:19 -0400)]
poky-qemu-ifup: allow multiple tap devices

This patch bases the tap IP address on the device number, providing
each device with its own IP address.

Signed-off-by: Jeff Dike <jdike@linux.intel.com>
14 years agoqemu: Use a TAP device instead of slirp for networking
Jeff Dike [Wed, 21 Jul 2010 15:04:59 +0000 (11:04 -0400)]
qemu: Use a TAP device instead of slirp for networking

With this patch, a persistent TAP device is set up by poky-qemu-ifup,
which is now run before qemu.  The qemu command line now uses the
device that was constructed (rather than the hard-coded tap0) and it
is told not to run any networking scripts.

When qemu shuts down, poky-qemu-ifdown removes the TAP device.

sudo use - sudo is used to run poky-qemu-ifup.  sudo is no longer used
to run qemu, as qemu no longer needs privileges to set up networking.
poky-qemu-ifdown is run without privileges, as you can remove a TAP
device which you own.

Signed-off-by: Jeff Dike <jdike@linux.intel.com>
14 years agounfs-server: new userspace nfs recipe
Scott Garman [Fri, 13 Aug 2010 04:02:58 +0000 (21:02 -0700)]
unfs-server: new userspace nfs recipe

This is a simple userspace NFS server, derived from one which was
previously used in openSUSE 10.x. Wind River contributed many of the
patches.

This package is not intended for target installations, only -native
and -nativesdk use.

Enabling nativesdk for readline, sqlite3, and pseudo was required, as
well as a few new autoconf siteconfig entries.

Signed-off-by: Scott Garman <scott.a.garman@intel.com>
14 years agoglibc: remove all old recipes
Nitin A Kamble [Thu, 19 Aug 2010 23:18:28 +0000 (16:18 -0700)]
glibc: remove all old recipes

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
14 years agovalgrind: upgrade for eglibc 2.12
Nitin A Kamble [Tue, 17 Aug 2010 23:31:18 +0000 (16:31 -0700)]
valgrind: upgrade for eglibc 2.12

the current version only works with libc upto 2.10
hence upgraded to support eglibc 2.12

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
14 years agogdb: disable installation of bfd/iberty/opcodes
Kevin Tian [Fri, 20 Aug 2010 05:13:36 +0000 (13:13 +0800)]
gdb: disable installation of bfd/iberty/opcodes

Those are assumed to be provided by binutils, so we disable installation
here. Without doing so, a broken libbfd.la from gdb is installed to sysroot
which overrides binutils version and then hurt other packages such as
oprofile

also fix --with-readline to --with-system-readline

Signed-off-by: Kevin Tian <kevin.tian@intel.com>
14 years agobitbake/utils.py: Allow copyfile to copy files which aren't readable
Richard Purdie [Fri, 20 Aug 2010 14:11:44 +0000 (15:11 +0100)]
bitbake/utils.py: Allow copyfile to copy files which aren't readable

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agobitbake/utils.py: Improve better_exec debug output
Richard Purdie [Fri, 20 Aug 2010 11:52:33 +0000 (12:52 +0100)]
bitbake/utils.py: Improve better_exec debug output

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agobitbake: Correctly route events from the worker to the server
Richard Purdie [Fri, 20 Aug 2010 11:25:19 +0000 (12:25 +0100)]
bitbake: Correctly route events from the worker to the server

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agobitbake/knotty: Exiting as soon as a fatal is seen is not desirable as the stacktrace...
Richard Purdie [Fri, 20 Aug 2010 11:24:13 +0000 (12:24 +0100)]
bitbake/knotty: Exiting as soon as a fatal is seen is not desirable as the stacktrace won't be seen

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agobitbake/cooker.py: Don't init the fetcher in the worker case
Richard Purdie [Fri, 20 Aug 2010 10:26:14 +0000 (11:26 +0100)]
bitbake/cooker.py: Don't init the fetcher in the worker case

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agobitbake/persist_data: Attempt to fix locking issues
Richard Purdie [Fri, 20 Aug 2010 10:24:02 +0000 (11:24 +0100)]
bitbake/persist_data: Attempt to fix locking issues

It appears the timeout sometimes has no effect and we see database access failures. Combat
this by wrapping the execute function in all cases and retrying manually ourselves.

Thanks to Kevin Tian for help debugging this.

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agobitbake/data.py: Handle exceptions in export_bars in the same way as emit_var()
Richard Purdie [Fri, 20 Aug 2010 08:40:49 +0000 (09:40 +0100)]
bitbake/data.py: Handle exceptions in export_bars in the same way as emit_var()

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agobitbake/server: Raise a shutdown event if we're seeing exceptions
Richard Purdie [Fri, 20 Aug 2010 08:38:24 +0000 (09:38 +0100)]
bitbake/server: Raise a shutdown event if we're seeing exceptions

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agobitbake/runqueue.py: Ensure the flush happens as late as possible to avoid data mixups
Richard Purdie [Thu, 19 Aug 2010 22:39:19 +0000 (23:39 +0100)]
bitbake/runqueue.py: Ensure the flush happens as late as possible to avoid data mixups

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agobitbake/utils.py: Give useful debug information when no lineno information is available
Richard Purdie [Thu, 19 Aug 2010 22:26:46 +0000 (23:26 +0100)]
bitbake/utils.py: Give useful debug information when no lineno information is available

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agoRename packaged-staging2 -> sstate and remove the old empty class
Richard Purdie [Thu, 19 Aug 2010 21:48:17 +0000 (22:48 +0100)]
Rename packaged-staging2 -> sstate and remove the old empty class

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agoRevert "base.bbclass: Temporarily disable fakeroot for install/package until other...
Richard Purdie [Thu, 19 Aug 2010 21:41:00 +0000 (22:41 +0100)]
Revert "base.bbclass: Temporarily disable fakeroot for install/package until other fixes get merged"

This reverts commit 2225e1214285f0e9a3c0ee2962b3d678c5c05292 and reenables
the functionality now its safe to do so.

14 years agoclasses: Disable setscene tasks for initial testing
Richard Purdie [Thu, 19 Aug 2010 19:08:53 +0000 (20:08 +0100)]
classes: Disable setscene tasks for initial testing

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agosanity.bbclass: Don't run the handler in worker context
Richard Purdie [Mon, 16 Aug 2010 15:41:43 +0000 (16:41 +0100)]
sanity.bbclass: Don't run the handler in worker context

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agopackage_*.bbclass: Always run these tasks under fakeroot
Richard Purdie [Mon, 16 Aug 2010 15:40:52 +0000 (16:40 +0100)]
package_*.bbclass: Always run these tasks under fakeroot

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agobitbake/runqueue.py: Fix runqueue UI issues
Richard Purdie [Thu, 19 Aug 2010 21:35:33 +0000 (22:35 +0100)]
bitbake/runqueue.py: Fix runqueue UI issues

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agobitbake: Switch to use subprocess for forking tasks and FAKEROOTENV to run shell...
Richard Purdie [Mon, 16 Aug 2010 15:37:29 +0000 (16:37 +0100)]
bitbake: Switch to use subprocess for forking tasks and FAKEROOTENV to run shell and python under a fakeroot environment

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agoRemove packaged-staging.bbclass
Richard Purdie [Wed, 11 Aug 2010 13:39:31 +0000 (14:39 +0100)]
Remove packaged-staging.bbclass

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agodeploy.bbclass: use new style staging for deploy tasks
Joshua Lock [Thu, 5 Aug 2010 13:25:20 +0000 (13:25 +0000)]
deploy.bbclass: use new style staging for deploy tasks

All tasks which implement a do_deploy should inherit this class to have the
changes in the deploy task staged.

Update recipes which include a do_deploy function to inherit this class and
to use DEPLOYDIR rather than DEPLOY_DIR_IMAGE.

Signed-off-by: Joshua Lock <josh@linux.intel.com>
14 years agopackage.bbclass: Implement packaged-staging2
Richard Purdie [Thu, 5 Aug 2010 13:16:59 +0000 (14:16 +0100)]
package.bbclass: Implement packaged-staging2

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agostaging.bbclass: Convert to packaged-staging
Richard Purdie [Thu, 5 Aug 2010 13:16:28 +0000 (14:16 +0100)]
staging.bbclass: Convert to packaged-staging

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
14 years agopackage_(deb|rpm): Update to use packaged-staging2
Joshua Lock [Thu, 5 Aug 2010 09:52:10 +0000 (10:52 +0100)]
package_(deb|rpm): Update to use packaged-staging2

Signed-off-by: Joshua Lock <josh@linux.intel.com>