Richard Purdie [Mon, 13 Jul 2015 11:04:52 +0000 (12:04 +0100)]
dpkg: Fix for Fedora22 and new versions of tar
They managed to 'break' tar. Again. Sorry, they fixed a regression
which broke dpkg-deb.
The addition of:
http://git.savannah.gnu.org/cgit/tar.git/commit/?id=163e96a0e619a900eab6de827c7c5749ecc9d3f2
("Bugfix: entries read from the -T file did not get proper matching_flag.")
means that the no-recursion option gets lost. This leads to many files getting included
multiple times, along with files which shouldn't be there.
The commit message is horrendous. The patch actually makes the option positional
(as documnted since 2003) and therefore doesn't affect the input from the -T option.
Moving the --no-reursion option to earlier in the command avoids the bug.
The bug was not present in tar 1.28 however it has been backported in at least
Fedora 22 and heading into Fedora 21.
Redhat reports of issue:
https://bugzilla.redhat.com/show_bug.cgi?id=1230762 [tar]
https://bugzilla.redhat.com/show_bug.cgi?id=1241508 [dpkg]
Discussion of bug in upstream tar:
http://www.mail-archive.com/bug-tar@gnu.org/msg04799.html
Signed-off-by: Jan Wetter <jan.wetter@mikrom.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org> Signed-off-by: Joshua Lock <joshua.lock@collabora.co.uk>
Signed-off-by: Roy Li <rongqing.li@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org> Signed-off-by: Joshua Lock <joshua.lock@collabora.co.uk>
Richard Purdie [Tue, 7 Jul 2015 12:22:15 +0000 (13:22 +0100)]
oeqa/bbtests: Fix race over DL_DIR and SSTATE_DIR
Running "-c cleanall" on shared DL_DIR and SSTATE_DIR is antisocial.
It leads to hard to debug races where we wonder why files disappear
and reappear from those directories.
Fix this by using a specific set of directories for these tests. This
avoids a long standing bug on the autobuilder where aspell and man
sources would disappear.
Mariano Lopez [Fri, 26 Jun 2015 05:56:00 +0000 (05:56 +0000)]
report-error.bbclass: Added file syncronization.
errorreport_handler would fail if several errors are
triggered at the same time because of two proccess
writting to the same file. This patch add the required
syncronization to handle concurrent process.
[YP #7899]
Signed-off-by: Mariano Lopez <mariano.lopez@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Saul Wold [Fri, 26 Jun 2015 18:23:06 +0000 (11:23 -0700)]
rootfs.py: Add check for kernel modules before running depmod
Add a check for kernel modules so we don't un-necessarily run the depmods, this
will also handle the case with linux-dummy does not place the kernel-abiversion
since it also does not have kernel modules.
[YOCTO #7884]
Signed-off-by: Saul Wold <sgw@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
gdb: Add missing runtime dependency for python option
With python enabled, gdb refuses to start without core
python modules:
| Could not find platform independent libraries <prefix>
| Could not find platform dependent libraries <exec_prefix>
| Consider setting $PYTHONHOME to <prefix>[:<exec_prefix>]
| ImportError: No module named site
Signed-off-by: Andreas Oberritter <obi@opendreambox.org> Signed-off-by: Andre McCurdy <armccurdy@gmail.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Kai Kang <kai.kang@windriver.com> Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com> Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Andre McCurdy [Wed, 27 May 2015 18:48:49 +0000 (11:48 -0700)]
gnutls: use pkg-config to locate zlib
AC_LIB_HAVE_LINKFLAGS can sometimes find host libs and is therefore not
robust when cross-compiling. Remove it for zlib and use PKG_CHECK_MODULES
instead.
Richard Purdie [Thu, 25 Jun 2015 15:09:14 +0000 (16:09 +0100)]
lib/devtool/standard: Fix patch cleanup
If patches fail to apply with git, quilt it used as a fallback. If that
happens, the code in this class is meant to handle cleanup of these patch
files. In the case where ${S} is a subdir of the git tree, the code doesn't
correctly set the patches directory.
This change correctly sets the patches directory (which is different to the
location of the git repository).
Richard Purdie [Thu, 25 Jun 2015 15:08:26 +0000 (16:08 +0100)]
lib/oe/patch: Fix git patch application for source in subdirectory
Similarly to:
http://git.yoctoproject.org/cgit.cgi/poky/commit/meta/lib/oe/patch.py?id=f205ccaf48ac36f4b26efc4aeb2e9d2939b28646
we need to fix patch application for source which is in a subdirectory.
Passing "." as the git directory or work-dir appears to work (or is ignored)
in some versions of git but does not work in others, probably quite correctly.
Since we have reporoot from the above patch, pass this in directly.
This bug caused this sanity test failure on some machines:
FAIL: test_devtool_modify_git (oeqa.selftest.devtool.DevtoolTests)
----------------------------------------------------------------------
Traceback (most recent call last):
File "/home/pokybuild/yocto-autobuilder/yocto-worker/nightly-oe-selftest/build/meta/lib/oeqa/selftest/devtool.py", line 390, in test_devtool_modify_git
self.assertEqual(result.output.strip(), "", 'Created git repo is not clean')
AssertionError: '?? util/mkelfImage/patches/' != '' : Created git repo is not clean
since git apply would fail, it would then fall back to quilt
and the git tree would be left unclean.
terminal.py: Allow devshell/menuconfig on recent gnome-terminal
Recent versions of gnome-terminal does not support non-UTF8 charset:
https://bugzilla.gnome.org/show_bug.cgi?id=732127 as a result, devshell and
menuconfig tasks silently hang (error found on trace log of 'strace -f -v
-s 8192 -e write=2 bitbake -c devshell quilt-native': "Non UTF-8 locale
(ANSI_X3.4-1968) is not supported!"). As a workaround, clearing the LC_ALL
environment variable so it uses the locale. Once fixed on the gnome-terminal
project, this should be removed.
Tested on gnome-terminal versions:
GNOME Terminal 3.4.1.1
GNOME Terminal 3.14.2 (Default on Ubuntu 15.04)
[YOCTO #7791]
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
we need to stop the preprocessor from generating the #line directives
or we run into issues like
| checking for apr_int64_t Python/C API format string...
| configure: error: failed to recognize APR_INT64_T_FMT on this platform
| Configure failed. The contents of all config.log files follows to aid
debugging
| ERROR: oe_runconf failed
Rightly subversion should be fixed but lets leave that to subversion
folks
runqemu-internal: set mutual exclusiveness for serial and nographic options
Use "-nographic" option only if "serial" option is not
specified. Otherwise we get below error when
'runqemu <kernel_image> <rootfs_image> serial' is executed,
(snip)
QEMU 2.2.0 monitor - type 'help' for more information
(qemu) qemu-system-aarch64: -serial stdio: cannot use stdio by multiple
character devices
-- CUT --
runqemu: fix MACHINE being detected as qemuarm for qemuarm64 kernel image
Basically, runqemu script autodetects MACHINE type based on the
kernel image name; if MACHINE name is not specified. Since 'qemuarm'
string is common in both qemuarm amnd qemuarm64 kernel image names, the
MACHINE type is considered as 'qemuarm' even when qemuarm64 kernel
image name is given in command line.
python3-ctypes: Fix cross compilation for arm targets
When cross compiling for arm targets ctypes compilation fails because
it uses _sysconfigdata from the HOST, this patches makes it use the
one from TARGET fixing compilation of this module
[YOCTO #7873]
Signed-off-by: Alejandro Hernandez <alejandro.hernandez@linux.intel.com> Signed-off-by: Jonas Göransson <jonas.goransson@qmatic.com>
Signed-off-by: Mark Hatle <mark.hatle@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Roy Li <rongqing.li@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
distutils: regenerate pyc files after being modified by sed
py files are edited by sed and therefore *.pyc files are recreated on first boot, but if you have a read-only filesystem this is not possible. This patch creates pyc files directly after the py files are modified.
Joshua Lock [Wed, 17 Jun 2015 08:33:18 +0000 (09:33 +0100)]
distutils-common-base: add to, don't set, FILES_${PN}
If we set FILES_${PN} and a recipe inherits other classes that
modify FILES_${PN} *before* distutils-common-base is included, any
changes to FILES_${PN} made by those classes are lost.
Instead, append the additional directories we want to include in
FILES_${PN}
Richard Purdie [Thu, 11 Jun 2015 22:03:28 +0000 (23:03 +0100)]
wayland: Fix hardlink corruption issue
The way this code was working, the m4 file is hardlinked to the
copies which would be packaged and could lead to the native m4
file being used in the target packages.
By removing the file first the hardlink is broken and this avoids
corruption (since cp uses open to change the file contents).
Signed-off-by: Yue Tao <Yue.Tao@windriver.com> Signed-off-by: Wenzong Fan <wenzong.fan@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
valgrind: add configure option to build only 32-bit or 64-bit components
Building both 32-bit and 64-bit binaries in valgrind at a time would
lead to following QA issue as below,
(snip)
ERROR: QA Issue: Architecture did not match (62 to 3) on ${WORKDIR}/valgrind/3.10.1-r0/packages-split/valgrind-dbg/usr/lib64/valgrind/.debug/vgpreload_exp-sgcheck-x86-linux.so
ERROR: QA Issue: Architecture did not match (62 to 3) on ${WORKDIR}/valgrind/3.10.1-r0/packages-split/valgrind-dbg/usr/lib64/valgrind/.debug/getoff-x86-linux
ERROR: QA Issue: Architecture did not match (62 to 3) on ${WORKDIR}/valgrind/3.10.1-r0/packages-split/valgrind-dbg/usr/lib64/valgrind/.debug/vgpreload_core-x86-linux.so
ERROR: QA Issue: Architecture did not match (62 to 3) on ${WORKDIR}/valgrind/3.10.1-r0/packages-split/valgrind-dbg/usr/lib64/valgrind/.debug/vgpreload_memcheck-x86-linux.so
-- CUT --
hence perform only one type of build 32-bit or 64-bit, but not both.
Robert Yang [Tue, 9 Jun 2015 06:46:25 +0000 (23:46 -0700)]
libpostproc: pass correct libdir
When MACHINE=qemux86-64 and enable multilib:
ERROR: QA Issue: libpostproc: Files/directories were installed but not shipped in any package:
/usr/lib
/usr/lib/libpostproc.so.52.3.0
/usr/lib/libpostproc.so
/usr/lib/libpostproc.so.52
/usr/lib/.debug
/usr/lib/.debug/libpostproc.so.52.3.0
Please set FILES such that these items are packaged. Alternatively if they are unneeded, avoid installing them or delete them within do_install. [installed-vs-shipped]
Pass the right libdir to configure as otherwise it assumes $prefix/lib
which may be wrong.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Lei Maohui <leimaohui@cn.fujitsu.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Dave Lerner [Fri, 22 May 2015 15:42:42 +0000 (10:42 -0500)]
valgrind: remove arm tests that don't compile
[Yocto #7453]
Corrects the original commit for the patch that removed ARM ptest CFLAGS
settings. Since the flags could be set by a user, the flags should
be kept in place during compilation. By keeping the original up-stream
CFLAGS for the tests, then additional tests successfully compile
for all tested ARM tunings.
However, there were still two tests listed below that did not compile
for any beaglebone tuning that is valid for valgrind. With the updated
patch, the set of excluded ARM ptests and their respective build
failures are:
intdiv - fails for all beaglebone tunings with 2 errors:
{standard input}:(40 or 41): Error: selected processor does not
support Thumb mode `udiv r3,r9,r10'
{standard input}:(72 or 73): Error: selected processor does not
support Thumb mode `sdiv r3,r9,r10'
vcvt_fixed_float_VFP - fails for all beaglebone tunings in one of
two ways:
with neon tuning (-mfpu=neon) fails with Internal Compiler Error
without neon tuning fails with 3 errors:
{standard input}:33: Error: selected FPU does not support
instruction -- `vcvt.f32.s32 s15,s15,#1'
{standard input}:58: Error: selected FPU does not support
instruction -- `vcvt.f32.s32 s15,s15,#32'
{standard input}:136: Error: selected FPU does not support
instruction -- `vcvt.f32.u32 s15,s15,#1'
After applying this commit, the valgrind ARM ptests compile without
errors for tunings:
armv7[t][hf][b][-neon] cortexa8[t][hf][-neon]
where the tuning [option] was successfully compiled, both with
and without the 'option', and in combination with all other options.
Signed-off-by: Dave Lerner <dave.lerner@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Yue Tao <Yue.Tao@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Mon, 1 Jun 2015 21:13:23 +0000 (22:13 +0100)]
base: Fix license checksum rebuild problems
"MACHINE=qemux86-64 bitbake init-ifupdown; MACHINE=genericx86-64 bitbake init-ifupdown"
shows a rebuild when it would be expected. The reason is a LIC_FILES_CHKSUM which
contains file://${WORKDIR}, an absolute path which doesn't exist in the first build
but does in the second, causing a signature change and a rebuild.
Fix the problem by ignoring any file:// url which resolves since TMPDIR for
license file dependency purposes.
Paul Eggleton [Thu, 7 May 2015 13:52:25 +0000 (14:52 +0100)]
oe-selftest: devtool: add a proper test to see if tap devices exist
Check up front in test_devtool_deploy_target whether the tap devices
exist and skip if not. If we don't do this we get a significantly less
comprehensible error via pexpect.
Paul Eggleton [Fri, 8 May 2015 14:26:04 +0000 (15:26 +0100)]
devtool: if workspace layer exists, still ensure it's in bblayers.conf
When we run devtool, if the workspace layer already exists but isn't in
bblayers.conf (perhaps because it was previously created but
subsequently removed from bblayers.conf by the user) then we should add
it and notify the user, otherwise devtool operations won't work.
If we execute an external command, we ought to prepare for the
possibility that it can fail and handle the failure appropriately. We
can especially expect this to happen when running bitbake in this
scenario. Ensure we return the appropriate exit code to the calling
process.
Paul Eggleton [Tue, 28 Apr 2015 11:25:30 +0000 (12:25 +0100)]
recipetool: ensure git clone is standalone when extracting
If -x is specified and the specified URI was a git repository, we need
to ensure that the resulting clone is a sandalone and not one that has
pointers into the temporary fetch location or DL_DIR (since the git
fetcher does a local clone with -s). Split out the code from devtool
that already does this for "devtool modify -x" and reuse that.
Paul Eggleton [Tue, 28 Apr 2015 11:22:36 +0000 (12:22 +0100)]
recipetool: avoid second-level subdir when extracting
When -x was specified, we were getting the normal unpack subdirectory
which we don't really want - if there's only one subdirectory unpacked
then we should effectively copy just it to the extraction path, not as a
subdirectory under it.
Paul Eggleton [Mon, 27 Apr 2015 09:53:21 +0000 (10:53 +0100)]
devtool: update-recipe: check if source tree is a git repository
If you've done "devtool add" (or "devtool modify" without -x) then it's
possible that the external source tree is not a git repository, so we
should handle that case here instead of printing a traceback.
Markus Lehtonen [Mon, 27 Apr 2015 09:53:14 +0000 (10:53 +0100)]
devtool: modify: implement --no-same-dir
This option can be used to have a separate build directory, in order to
keep the srctree directory clean for packages that do not need to be
built in the source directory.
Markus Lehtonen [Mon, 27 Apr 2015 09:53:13 +0000 (10:53 +0100)]
devtool: modify: use B=S if that is the default for the recipe
Makes the build succeed for packages which do not support separate build
directory, e.g. zlib. The same outcome could be achieved with the
--same-dir option, but, it's generally hard to tell if a random package
would need that option. The negative side effect of this patch is that
dev srctree (of some packages that build fine without this modification)
gets dirtied by build artefacts.
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com> Acked-by: Paul Eggleton <paul.eggleton@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Roy Li <rongqing.li@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Note that the fix is only required when glibc is built for i686/multiarch,
so is not applicable in the default oe-core x86 configuration (which builds
glibc for i586 and therefore does not include SSE2 optimised memcpy etc).
Signed-off-by: Andre McCurdy <armccurdy@gmail.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Saul Wold [Tue, 12 May 2015 19:02:46 +0000 (12:02 -0700)]
lttng-modules: Update to stable version 2.6.1
This fixes a build issue with the 3.19.5 kernel where the regmap prototypes
have changed. The patch is rebased do to changes in the new version of the
Makefile.
Robert Yang [Thu, 28 May 2015 09:22:24 +0000 (02:22 -0700)]
python3: fix build for shared object
Fixed when build on armv7a_vfp_neon:
Python-3.3.3/Modules/_struct.o: relocation R_ARM_MOVW_ABS_NC against `a local symbol' can not be used when making a shared object; recompile with -fPIC
All the archs should use -fPIC when build shared object for linux.
Signed-off-by: Kai Kang <kai.kang@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org> Signed-off-by: Joshua Lock <joshua.lock@collabora.co.uk>
Signed-off-by: Kai Kang <kai.kang@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org> Signed-off-by: Joshua Lock <joshua.lock@collabora.co.uk>
Buffer overflow in the rc_mksid function in plugins/radius/util.c in
Paul's PPP Package (ppp) 2.4.6 and earlier, when the PID for pppd is
greater than 65535, allows remote attackers to cause a denial of
service (crash) via a start accounting message to the RADIUS server.
oe-core is using ppp 2.4.7, and this CVE say ppp 2.4.7 was not
effected, but I found this buggy codes are same between 2.4.6 and
2.4.7, and 2.4.7 should have this issue.
Signed-off-by: Kai Kang <kai.kang@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org> Signed-off-by: Joshua Lock <joshua.lock@collabora.co.uk>
Signed-off-by: Roy Li <rongqing.li@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Andre McCurdy <armccurdy@gmail.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Andre McCurdy <armccurdy@gmail.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Peter Seebach [Thu, 21 May 2015 20:36:35 +0000 (15:36 -0500)]
rootfs.py: Make set -x not break rootfs creation
The rootfs logfile check errors out if it finds an "exit 1" in
a log. But consider the shell idiom:
if ! $command; then exit 1; fi
and consider a postinstall script with a "set -x" for debugging
purposes (to get log output in /var/log/postinstall.log.)
Solution: Ignore lines prefixed with a +, because those show shell
code even if a specific fragment won't be executed.
Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Wed, 20 May 2015 14:38:49 +0000 (15:38 +0100)]
zlib: clean up base_libdir symlink
libz.so symbolic link created in ${libdir} is
../../${base_libdir}/libz.so.1.2.8. This doesn't work if base_libdir or libdir
is changed, so use oe.path.relative to construct the correct path at build time.
André Draszik [Mon, 18 May 2015 08:06:01 +0000 (09:06 +0100)]
icu-native: fix hard paths in native icu
The icu recipe installs icu-native twice, once in the usual location,
and once for cross builds into the path given by --with-cross-build.
This latter path is not included in the list of paths recognised by
chrpath.bbclass, hence the binaries in there retain the rpath as used
during compilation. This causes the package to not be relocateable
from sstate
[YOCTO #6851]
We use the infrastructure that is in place already, and simply set
PREPROCESS_RELOCATE_DIRS as necessary, and things start to work.
A buffer overflow flaw was found in the way glibc's gethostbyname_r() and
other related functions computed the size of a buffer when passed a
misaligned buffer as input. An attacker able to make an application call
any of these functions with a misaligned buffer could use this flaw to
crash the application or, potentially, execute arbitrary code with the
permissions of the user running the application.
Signed-off-by: Haris Okanovic <haris.okanovic@ni.com> Reviewed-by: Ben Shelton <ben.shelton@ni.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
native.bbclass: avoid unintended substring replacement when setting PROVIDES
The way native_virtclass_handler was implemented leaded to
unintended substring replacements when setting PROVIDES for
native providers, in case the original PROVIDES value contains
providees with common substrings.
Here's a practical case where the old behavior was problematic:
the oracle-jse-jdk-x86-64 recipe provides both virtual/java and
virtual/javac:
Signed-off-by: Mario Domenech Goulart <mario@ossystems.com.br> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Otavio Salvador [Mon, 11 May 2015 11:26:47 +0000 (08:26 -0300)]
gstreamer1.0-plugins-bad: Move EGL requirement for Wayland
The EGL is used by the Wayland backend. When building using Software
Rendering and without Wayland support the EGL backend is not available
so we should not require EGL for GLES2 support.
This fixes following build error:
,----
| ...
| checking for bcm_host_init in -lbcm_host... no
| checking for WAYLAND_EGL... no
| configure: error: Could not find the required EGL libraries
| Configure failed. The contents of all config.log files follows to aid debugging
| ERROR: oe_runconf failed
`----
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Martin Jansa [Thu, 14 May 2015 11:41:26 +0000 (13:41 +0200)]
boost: properly fix do_boostconfig re-execution
* it was partially fixed in:
commit 291e20a51544c640d07767d1dc32d762f4370f41
Author: Venkata ramana gollamudi <ramana.gollamudi@huawei.com>
Date: Fri Apr 13 11:42:46 2012 +0000
Subject: boost: fix re-execution of task
but with disadvantage that when CXX or *FLAGS variables were changed
it was continuing to use old values
* just remove the line before appending it with current values to fix
that
Martin Jansa [Tue, 28 Apr 2015 22:46:18 +0000 (00:46 +0200)]
e2fsprogs: install populate-extfs.sh
* install populate-extfs.sh from contrib, be aware that in order
to use it you need to set DEBUGFS shell variable, otherwise it will
try to use debugfs from relative path which is almost always
incorrect:
CONTRIB_DIR=$(dirname $(readlink -f $0))
DEBUGFS="$CONTRIB_DIR/../debugfs/debugfs"
Signed-off-by: Robert Yang <liezhi.yang@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org> Signed-off-by: Joshua Lock <joshua.lock@collabora.co.uk>
Richard Purdie [Wed, 13 May 2015 08:08:09 +0000 (09:08 +0100)]
glibc: Fix x32 make race
On x32 builds, sysd-syscalls appears malformed since the make-target-directory
appears on the wrong line. This causes races during the build process where you can
see failures like:
Assembler messages:
Fatal error: can't create [...]glibc/2.21-r0/build-x86_64-poky-linux-gnux32/time/gettimeofday.os: No such file or directory
Assembler messages:
Fatal error: can't create [...]glibc/2.21-r0/build-x86_64-poky-linux-gnux32/time/time.os: No such file or directory
The issue is that the carriage return is being escaped when it should
not be. The change to sysd-syscalls with this change:
which ensures the target directory is correctly created. Only x32 uses the vdso
code which contains the bug which is why the error only really appears on x32.
For recipes with PACKAGES_remove = "${PN}", the find which removes
.la files can race against deletion of other directories in WORKDIR
e.g.:
find: '/home/autobuilder/yocto-autobuilder/yocto-worker/nightly-oe-selftest/build/build/tmp/work/qemux86_64-poky-linux/init-ifupdown/1.0-r7/sstate-build-populate_lic': No such file or directory
| WARNING: /home/autobuilder/yocto-autobuilder/yocto-worker/nightly-oe-selftest/build/build/tmp/work/qemux86_64-poky-linux/init-ifupdown/1.0-r7/temp/run.do_configure.6558:1 exit 1 from
| find /home/autobuilder/yocto-autobuilder/yocto-worker/nightly-oe-selftest/build/build/tmp/work/qemux86_64-poky-linux/init-ifupdown/1.0-r7 -name \*.la -delete
Signed-off-by: Robert Yang <liezhi.yang@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
unzip 6.0 allows remote attackers to cause a denial of service
(out-of-bounds read or write and crash) via an extra field with
an uncompressed size smaller than the compressed field size in a
zip archive that advertises STORED method compression.
Buffer overflow in the charset_to_intern function in unix/unix.c in
Info-Zip UnZip 6.10b allows remote attackers to execute arbitrary code
via a crafted string, as demonstrated by converting a string from CP866
to UTF-8.
Signed-off-by: Roy Li <rongqing.li@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Bruce Ashfield [Fri, 8 May 2015 03:36:15 +0000 (23:36 -0400)]
kernel-yocto: propagate in tree defconfigs to WORKDIR
As reported by Steffen Pankratz <Steffen.Pankratz@elektrobit.com>, the
previous logic of KBUILD_DEFCONFIG processing would not propagate an in
tree defcofig to WORKDIR if one was not already present.
We fix the propagation by copying the in tee config if a defconfig is
not already in WORKDIR.
Additionally we only warn (versus copying) if an in tree configuration
is specified, is different than the WORKDIR version and isn't copied.
Bruce Ashfield [Fri, 8 May 2015 03:36:14 +0000 (23:36 -0400)]
linux-yocto: fix race between checkout and meta data generation
There are two tasks that must run before a linux-yocto kernel is built.
- Kernel checkout and relocation to work-shared (kernel_checkout)
- Meta data gathering and configuration prep (kernel_metadata)
The current task definitions for both are simply "before do_patch",
which is correct, but kernel_checkout must run before and not race with
kernel_metadata.
So we set the definition of kernel_checkout to be more specific and
enforce the proper ordering.
