When run valgrind, following error appears:
==2254== FATAL: can't open suppressions file "/usr/lib/valgrind/default.supp"
(From OE-Core rev: 0b3261d513cdad80174a9b9e804981c50bcb7ca2)
(From OE-Core rev: 95756cfbb7a9348b23cb46a49a5509e57e973faf)
Signed-off-by: Zhenhua Luo <b19537@freescale.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
the builddir is put in front of the LD_LIBRARY_PATH, causing dynamically
linking of target library with native tclsh.
Fix this behavior to cross build tcl correctly.
This issue got exposed when eglibc-2.15 was configured for the target.
(From OE-Core rev: 8e25fe0ecc3d6fe2d5456b525c5014554bc70cfe)
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
While building libtool, the libtool script itself will be regenerated
because OE modifies a dependency[1]. With -jX, this operation (-->
removal, creation of non-x file, 'chmod a+x') can happen at a time when
the script is going to be executed. This can cause errors like:
| arm-linux-gnueabi-libtool: compile: ccache arm-linux-gnueabi-gcc ...
| ...
| /bin/sh ./config.status libtool
| ...
| arm-linux-gnueabi-libtool: compile: ccache arm-linux-gnueabi-gcc ...
| /bin/sh: ./arm-linux-gnueabi-libtool: Permission denied
| make[2]: *** [libltdl/libltdl_libltdl_la-lt__alloc.lo] Error 126
I am not sure whether the custom do_compile_prepend() is still needed.
For now only the issue above will be fixed by executing ./config.status
yet again.
[1] see 648290d5bf
(From OE-Core rev: 15204a6cbcdbbb84e02da05b1fb15644fe7df332)
Signed-off-by: Enrico Scholz <enrico.scholz@sigma-chemnitz.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This fixes the issue where gcc invokes the linker with an incorrect -L
library location and gives up because it can't find libraries. It was
looking in a /lib folder instead of /lib64
(From OE-Core rev: aa010039a38188f1b1b38a978287d1597138b8b9)
Signed-off-by: Matthew McClintock <msm@freescale.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
MakeMaker has a bug where it does not propagate CC/LD/etc information
down to subproject it generates Makefiles for... this recipe has has an
Expat subproject which has issues building if we are using sstate-cache
and it will reference the old sysroots and be unable to build properly.
There is an upstream MakeMaker bug for this issue but we can work around
it by fixing up the Makefiles for now
See:
https://rt.cpan.org/Public/Bug/Display.html?id=28632
(From OE-Core rev: e1609123a6ca6aef18e48afe0ce61325da910fc1)
Signed-off-by: Matthew McClintock <msm@freescale.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
On Fedora 17, bash has moved to /usr/bin/bash and the configure process finds it
on the host machine there, this ensures that it is set correctly for the target.
[YOCTO #2363]
(From OE-Core rev: d54ff1f79f05ba5bd0e1006545e7f1e699998668)
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Reworked commit to fix merge conflicts with denzil branch.
Signed-off-by: Scott Garman <scott.a.garman@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Currently since configure.in in is in a subdirectory, we don't reautoconf the
recipe. We really need to do this, to update things like the libtool script used
and fix various issues such as those that could creep in if a reautoconf is
triggered for some reason. Since this source only calls AM_INIT_AUTOMAKE to gain the
PACKAGE and VERSION definitions and that macro now errors if Makefile.am doesn't
exist, we need to add these definitions manually.
These changes avoid failures like:
----
| ...
| DssslApp.cxx:117:36: error: 'PACKAGE' was not declared in this scope
| DssslApp.cxx:118:36: error: 'VERSION' was not declared in this scope
| make[2]: *** [DssslApp.lo] Error 1
----
(From OE-Core rev: 87753615435c8aec7df5964045e24f13877cd7cc)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There is a bug if we:
1) bitbake diffutils with MACHINE=crownbay
2) bitbake diffutils with MACHINE=qemux86
3) bitbake core-image-sato with MACHINE=crownbay
Then the diffutils.i586 would be installed to the crownbay's image, this
is because diffutils.i586 is newer than diffutils.core2, and rpm doesn't
respect to the arch priorities:
We have put the archs in order in _solve_dbpath:
crownbay/solvedb:core2/solvedb:i586/solvedb:all/solvedb
Fix rpm to respect to the order, for example, if it finds a pkg in both
core2/ and i586/, and the core2/ comes first, it should not use the one
in i586/ even if it's build time is newer.
Note: Don't worry about the _free(*ptr), it can check whether ptr is
NULL or not.
This is for the denzil branch, and the master branch also needs it.
[YOCTO #2360]
(From OE-Core rev: 2199e6b9c82bb2b6738e87903f30329586db20e2)
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
While building on distros like fedora17, which has /bin/perl,
the target perl scripts get perl path also as /bin/perl.
And that is not correction path of perl on the target.
This commit avoids this error.
| error: Failed dependencies:
| /bin/perl is needed by quilt-0.51-r2.i586
NOTE: package core-image-sato-sdk-1.0-r0: task do_rootfs: Failed
ERROR: Task 8
(/home/nitin/prj/poky.git/meta/recipes-sato/images/core-image-sato-sdk.bb,
do_rootfs) failed with exit code '1'
(From OE-Core rev: c8c394bd806978c867f2fe82e4bde65c98764880)
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
python should search for db.h in inc_dirs and not in a hardcoded path.
If db.h is found but HASHVERSION is not 2 we avoid a warning by not.
adding this module to missing variable.
[YOCTO #1937]
(From OE-Core rev: 8eb3e52d39147f8cb98ec95857be17db0444098e)
Signed-off-by: Andrei Gherzan <andrei@gherzan.ro>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This patch was added for 64bit host machines. In the compile process python
is checking if platform is a 64bit platform using sys.maxint which is the host's
value. The patch fixes this issue so that python would check if TARGET machine
is 64bit not the HOST machine. In this way will have "dl" and "imageop" modules
built if HOST machine is 64bit but the target machine is 32bit.
[YOCTO #1937]
(From OE-Core rev: 22ae3959f40845ebcc00413ccf733539472a1a81)
Signed-off-by: Andrei Gherzan <andrei@gherzan.ro>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Backport a fix from 1.16.x upstream to use fd instead of stream-based
I/O in dpkg-deb, which avoids the use of fflush() on an input stream
(the behaviour of which is undefined by POSIX, and appears to have
changed in the version of glibc introduced in Fedora 16 and presumably
other systems).
Fixes [YOCTO #1858].
(From OE-Core rev: b1c28667592e736115ab5e603a12c2723b939cf2)
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
patch.bbclass orignally pointed at /usr/bin/quiltrc for an empty
version to ensure that no user setting were picked up, change this
to /etc/quiltrc in the Native sysroot since we now have a native
sysconfdir.
Make sure that the quiltrc is actually installed in the Native
sysconfdir, not the target, so fix this after the recipe split.
(From OE-Core rev: aec4cdc6efda430a0965d6b3b4f84c7943390273)
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
WARNING: For recipe gcc, the following files/directories were installed but not shipped in any package:
WARNING: /usr/include
WARNING: /usr/lib/gcc/arm-poky-linux-gnueabi/4.6.4/plugin/libgcc
WARNING: /usr/lib/gcc/arm-poky-linux-gnueabi/4.6.4/plugin/libgcc/config
WARNING: /usr/lib/gcc/arm-poky-linux-gnueabi/4.6.4/plugin/libgcc/config/arm
WARNING: /usr/lib/gcc/arm-poky-linux-gnueabi/4.6.4/plugin/libgcc/config/arm/bpabi-lib.h
(From OE-Core rev: cf49cf3958b24fdb89d57abbf1f1b30c07a06030)
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
WARNING: For recipe libgcc, the following files/directories were installed but not shipped in any package:
WARNING: /usr/lib/powerpc-poky-linux/4.6.4/ecrti.o
WARNING: /usr/lib/powerpc-poky-linux/4.6.4/ncrti.o
WARNING: /usr/lib/powerpc-poky-linux/4.6.4/ecrtn.o
WARNING: /usr/lib/powerpc-poky-linux/4.6.4/ncrtn.o
(From OE-Core rev: 580d734ddc928aaaac9acaa248427b01731074f2)
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
WARNING: For recipe binutils, the following files/directories were installed but not shipped in any package:
WARNING: /usr/bin/embedspu
(From OE-Core rev: 15c8ea4d35edbcaf03c94aba06ded85851679157)
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Bump PR value due to the commit
c6c701f424aeb502d20ff02d02712e56f4e259a5.
(From OE-Core rev: b6ee2880fccf04923ede31256ea418451cbf2e46)
Signed-off-by: Lianhao Lu <lianhao.lu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The move of libcrypto to /lib instead of /usr/lib has broken the _hashlib module
compilation. There were also a number of other failing modules which should
have been building correctly. This turned out partly to be the /lib issue
but also due to a number of native paths creeping into compiler commandlines.
These changes add in /lib as part of the searh directory and remove
a number of host contamination issues within setup.py. Post release we
should really further go through this file and just delete large sections
of it as its hard to be sure what strange paths python is injecting as
search paths.
This patch also fixes issues where re-execution of the compile task
would corrupt the Makefile in various ways, again leading to puzzling
paths within the configuration.
(From OE-Core rev: 20e2761e1da1cb5dcd267e161f2a6b6a429e9f39)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The -nativesdk pseudo wrapper setting LD_LIBRARY_PATH turned out to be a
bad idea since it can mix up different libc and lib-dl verisons which
may or may not work depending on the phase of the moon.
As an alternative to solving the original problem, this patch drops the
symbol version requirement on memcpy which allows pseudo to work with
libc's back to 2.7 which should be sufficient for our supported targets
using nativesdk.
[YOCTO #2299]
[YOCTO #2351]
(From OE-Core rev: c6c701f424aeb502d20ff02d02712e56f4e259a5)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This fix addresses various issues seen in qemux86-64 images:
* scroll bars in matchbox-terminal not working
* files not appearing in pcmanfm
* warnings on the console from glib/gobject about invalid gdouble values
Its due to an emulation issue in qemu which the backported patch fixes.
I managed to debug it to a specific function, Khem found the qemu patch
to backport, thanks Khem!
[YOCTO #1906]
(From OE-Core rev: 69d083f8b8d8f7d095ed5682d305870c4d93fe62)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There is an error when build meta-toolchain-sdk on Ubuntu 10.10:
| Installing NATIVESDK packages
| Processing task-sdk-host-nativesdk...
| Processing task-cross-canadian-mips...
| error: Failed dependencies:
| libcurl.so.4 is needed by qemu-nativesdk-0.15.1-r5.i686
NOTE: package meta-toolchain-gmae-1.0-r7: task do_populate_sdk: Failed
ERROR: Task 21 (/buildarea2/lyang1/poky/meta/recipes-core/meta/meta-toolchain-gmae.bb, do_populate_sdk) failed with exit code '1'
This is because qemu would depend on curl if it happened to be
available at build time and could confuse the -native and -nativesdk
librbaries.
[YOCTO #2305]
(From OE-Core rev: 566ad643efd03db34b59a427f48ca55ce613cf89)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
After building perl package, re-execution of compile task recursively
substitutes the path, making it an invalid path.
Fixed to prevent recursive substitution.
Similar case as [Yocto #2194]
(From OE-Core rev: 16542d982d86d42d3189d47a8180f0f71646a9ca)
Signed-off-by: Venkata Ramana Gollamudi <ramana.gollamudi@huawei.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The generated parser had warnings regarding signess and return check
which makes Linux Kernel's perf tool from 3.4 release to fail without
those patches.
(From OE-Core rev: f3d7197252d1ede627a561fbd5b3b7fb759bf75b)
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Generally distros keep perl at /usr/bin/perl
Fedora 17 also has /bin/perl
this causes automake-nativesdk build on such distros to put perl
interpreter path in the perl scripts as /bin/perl
But we set perl location for target as /usr/bin/perl
This mismatch of perl path causes failure of rootfs image creation
like this:
| error: Failed dependencies:
| /bin/perl is needed by automake-nativesdk-1.11.2-r2.x86_64
NOTE: package meta-toolchain-gmae-1.0-r7: task do_populate_sdk: Failed
The second sed command is for such a case:
eval 'case $# in 0) exec /bin/perl -S "$0";; *) exec /bin/perl -S "$0" "$@";; esac'
This line has two "/bin/perl" and we can't use a line number to tell
sed which line it is since the line numbers in different files are
defferent.
[YOCTO #2304]
(From OE-Core rev: ee7a86c06c2289f01aa5f1da958ce51523495572)
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Generally distros keep perl at /usr/bin/perl
Fedora 17 also has /bin/perl
this causes autoconf-nativesdk build on such distros to put perl
interpreter path in the perl scripts as /bin/perl
But we set perl location for target as /usr/bin/perl
This mismatch of perl path causes failure of rootfs image creation
like this:
| error: Failed dependencies:
| /bin/perl is needed by autoconf-nativesdk-2.68-r6.x86_64.rpm
NOTE: package meta-toolchain-gmae-1.0-r7: task do_populate_sdk: Failed
The second sed command is for such a case:
eval 'case $# in 0) exec /bin/perl -S "$0";; *) exec /bin/perl -S "$0" "$@";; esac'
This line has two "/bin/perl" and we can't use a line number to tell
sed which line it is since the line numbers in different files are
defferent.
[YOCTO #2304]
(From OE-Core rev: 393908e53b243b16ab984da7f073be371e062946)
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The perl recipe's do_configure() does some manual sed replacements over
the entire WORKDIR. Fix the following issues with this:
* Skip patches, which fixes re-execution of do_patch after do_configure
has run once
* Ensure that the replacement operation does nothing if do_configure is
re-executed
* Avoid unnecessarily modifying /usr/include paths within documentation
that will end up being packaged
* Fix a quoting issue in the expression used in the grep command that
ended up causing files that did not contain .*\.h after /usr/include
to be matched and modified.
The files modified during do_configure have been compared before and
after this patch to ensure there are no unexpected changes. Some
/usr/include paths that are not within documentation are no longer
being substituted however these are all within comments or scripts for
other Unix operating systems that are not applicable.
[YOCTO #2194]
(From OE-Core rev: 19255032e7744fce5cbe466e4869ded378d3b4f5)
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This was found to cause issues on the Yocto autobuilders and fixes do_compile
failures when guile-native has been relocated.
(From OE-Core rev: d928e91a57e6e9dfa6a7d4e888e1e1064d7fc668)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We need consistent configuration files for opkg so we might as well always use
the full list. This is equal to PACKAGE_ARCHS in the non-multilib case.
This fixes various multilib failures with ipk multilibs.
(From OE-Core rev: bc85abc5013d0c831cc3c3823df45536c293aaba)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
If the do_unpack task is re-executed, the sed_done stamp was not removed, the
sed commands wouldn't re-run and the build would fail in do_compile. This
patch ensures the stamp is in ${S} and that we clean that directory
when unpack runs so it is a build from sctatch at that point.
[YOCTO #2194]
(From OE-Core rev: 2b190155d74d4369c61d26b1f4f0c3cce389ddbc)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The sed command run during do_compile isn't idempotent. This change fixes it
so that it is and the compile task can be re-executed.
[YOCTO #2194]
(From OE-Core rev: c9dcfbd87d749baa0a4948738055193982aad26d)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There was a patch patching both configure and configure.in. Since the recipes
all autoreconf, this is pointless and means subsequent patch tasks fail.
The fix is to only patch configure.in.
[YOCTO #2194]
(From OE-Core rev: 1d8c40d5f2119b4c4ae8a9b2645ae572ea9a54ff)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There were puzzling failures when you make a force recompile of any gcc
component. The error was in do_configure with cross-compilation badness
being detected in config.log files.
gcc is different in that many of the config.log files are generated during
the do_compile phase. This means this host contamination issue has always
been present but only shows up on a rebuild.
The fix is to force the appropriate configuration variable to "none required"
then gcc won't look in the bad locations.
[YOCTO #2279]
(From OE-Core rev: e7135f14f3ef37f93f5c28b319464f3d6dd9b8da)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There can be a conflict between the nativesdk libc and the host system's
libc. It is assumed the nativesdk version is of an equal or higher version.
This is a particular issue for pseudo if its loading a system binary
since the system's libc might be used of an older verison which would
then confuse libpseudo.so when loaded as a preload.
To avoid this, set LD_LIBRARY_PATH so the nativesdk libc is always
used.
Since we now use --without-rpath, we can remove the MAKEOPTS RPATH workaround.
[YOCTO #2299]
(From OE-Core rev: a481fe3b9883aa744be3253e2b4b27e6e46eb059)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When guile-native was relocated, there were messages like:
;;; note: source file /srv/home/pokybuild/yocto-autobuilder/yocto-slave/emenlow/build/build/tmp/sysroots/x86_64-linux/usr/share/guile/2.0/ice-9/eval.scm
;;; newer than compiled /srv/home/pokybuild/yocto-autobuilder/yocto-slave/fri2/build/build/tmp/sysroots/x86_64-linux/usr/lib/guile/2.0/ccache/ice-9/eval.go
and this confuses things like the autogen-native "guile --version"
check. This patch ensures the wrapper script sets the necessary
variables correctly.
(From OE-Core rev: 7a254d8294cfbe4b717d083c3d7fcc5515a5ab4f)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This module is needed by python scripts from createrepo native package.
[YOCTO #2134]
RP: Add missing DEPENDS on python-native
(From OE-Core rev: 17f3cc1be0e8648cb4579793a9eb5a701deac85f)
Signed-off-by: Andrei Gherzan <andrei@gherzan.ro>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This is done because of this change in gdbm:
"gdbm: Package compat libs in gdbm-compat"
(From OE-Core rev: b91d8a07f736b7698650d25609245c89e0ed73d5)
Signed-off-by: Andrei Gherzan <andrei@gherzan.ro>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This brings the version comparision within opkg-utils into sync with that
in opkg itself.
[YOCTO #2233]
(From OE-Core rev: 1964be5432c1c378e80d7f1dec0999a2be71853b)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
If the toolchain is reused from sstate and ccache is installed, build failures
were occuring due to gcc trying to access the original sysroot rather than the
new one, particularly if the old sysroot existed but was not readable by the
current user.
This turns out of the an issue inside gcc to do with preservation of the sysroot
option. See the gcc patch for more details. It only triggers when preprocessed
sources are used which happens when ccache is used.
The same issue occurs with c++ and c++-cpp-output so the same fix is applied there.
[YOCTO #2074]
(From OE-Core rev: d3a3e81869631ba69874f6fc172240e3aac388f5)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
If this value is not set to 1, then systems with some 32-bit libraries
but no 32-bit version of libgcc installed will have pseudo-native fail
at do_compile. It should only really be set to 0 by those who know what
they are doing.
(From OE-Core rev: 489a36d3d6b67d706f5918638e1fbc05ccd59e21)
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Many of the standard C++ library include files have been edited to no
longer include <unistd.h> to remove namespace pollution in gcc 4.7, so
we need include <unistd.h> manually sometimes.
More information:
http://gcc.gnu.org/gcc-4.7/porting_to.html
[YOCTO #2278]
(From OE-Core rev: 5e3c12313edc4b51ddd432ecbbd3b1f5295b3c14)
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
[Updated patch header - sgw]
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
In lib/rpmts.c we need to make sure to look for missing dependencies,
such as /bin/sh, in both the providename and filepaths DB. Previously
the system would key off an initial '/' and only look in the filepaths DB.
This causes problems when a package such as "bash", has a Provides: /bin/sh.
(From OE-Core rev: 08fb526c42e85c89135960344745e878c5fb633a)
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
[Yocto #2251]
Add --without-rpath to avoid embedding rpaths into the pseudo
components.
(From OE-Core rev: ae978e9671fdbcb31e306308bfb816b4bd2b2496)
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
