If used with some packages using kconfig mechanism, the diffconfig
command generates wrong output format. Diff provides all options to
format the output correctly. This method formats as intended, is more
robust and works with the merge_config.h script from yocto-kernel-tools.
Li Wang [Mon, 28 Jul 2014 06:50:42 +0000 (02:50 -0400)]
nss: CVE-2013-5606
the patch comes from:
http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2013-5606
https://bugzilla.mozilla.org/show_bug.cgi?id=910438
http://hg.mozilla.org/projects/nss/rev/d29898e0981c
The CERT_VerifyCert function in lib/certhigh/certvfy.c in
Mozilla Network Security Services (NSS) 3.15 before 3.15.3 provides
an unexpected return value for an incompatible key-usage certificate
when the CERTVerifyLog argument is valid, which might allow remote
attackers to bypass intended access restrictions via a crafted certificate.
Shan Hai [Mon, 28 Jul 2014 05:18:50 +0000 (01:18 -0400)]
pulseaudio: fix CVE-2014-3970
The pa_rtp_recv function in modules/rtp/rtp.c in the module-rtp-recv module
in PulseAudio 5.0 and earlier allows remote attackers to cause a denial of
service (assertion failure and abort) via an empty UDP packet.
Fix it by picking a patch from pulseaudio upstream code.
Paul Gortmaker [Mon, 28 Jul 2014 02:58:36 +0000 (22:58 -0400)]
base-files: fix up misleading dir grouping names for lsb dirs
In commit b0df35f47fb79dc149504bd66d1186b3276f9510
("base-files: Add to make some directories needed by LSB.") there
were new groupings of dirs made. But rather than keep the grouping
names permissions based as they clearly were, it seems that it was
unfortunately misinterpreted as a counter, simply incrementing as
a new group was added.
This leads to dir3755 getting chmod'd to 0755 and dir4775 being
chmod'd to 2775 which to any new reader of the code is terribly
confusing.
Choose names that clearly reflect the permissions, and add a lsb
suffix to indicate the role.
Also note that the settings for /var/mail seem incorrect, and so
they have been aligned with what is seen in most common distros.
Paul Gortmaker [Mon, 28 Jul 2014 02:58:35 +0000 (22:58 -0400)]
base-files: remove strange 2775 perms from dirs like /home
Mark Hatle did the historical research into why this was so,
and found that it originally was via debian import, and then
it morphed over time:
<fray> +# chown -R root:root ${D}
<fray> +# cd ${D} && chown root:src usr/src
<fray> +# cd ${D} && chown root:staff var/local
<fray> +# cd ${D} && chown root:staff home
<fray> + cd ${D} && chmod 755 `find . -type d`
<fray> + cd ${D} && chmod 1777 `cat ${S}/debian/1777-dirs`
<fray> + cd ${D} && chmod 2775 `cat ${S}/debian/2775-dirs`
<fray> that was the original code
<fray> so THAT is why they're 2775..
<fray> they were original 'src' and 'staff'.. which has gotten lost
sometime in the last 10 years
<fray> AND the origin:
<fray> commit b45c9ed40bb4f893f99127a21776aef3ae888ad7
<fray> Author: Chris Larson <clarson@kergoth.com>
<fray> Date: Tue Sep 30 16:30:41 2003 +0000
<fray> Add base-files 3.0.10 (from debian).
<fray> thats where the brain damage started ... Debian in 2003..
Here we simply convert them to being normal 755 dirs.
Since 2004 we've been carrying a patch[1] make autoreconf pass --foreign to
automake. Presumably at the time this was due to many upstreams using
hand-coded bootstrap scripts that passed --foreign manually, but we were using
autoreconf. These days many projects have added foreign to AM_INIT_AUTOMAKE and
use autoreconf directly, so this patch isn't as critical as it used to be.
Richard Purdie [Sat, 26 Jul 2014 09:13:21 +0000 (10:13 +0100)]
populate_sdk_deb: Fix non x86_64 SDK builds
If building with SDKMACHINE=i686 and using the deb backend, populate_sdk
would fail. Its clear when looking at the options that the 32 bit values
were overwritten.
Replace this code with code similar to that used in package_deb itself.
Richard Purdie [Sat, 26 Jul 2014 08:01:43 +0000 (09:01 +0100)]
populate_sdk_base: Include do_rootfs depends for do_populate_sdk
"bitbake core-image-minimal -c populate_sdk; rm tmp -rf; bitbake core-image-minimal -c populate_sdk"
breaks due to missing dependencies on things like update-rc.d. Since
we're effecitvely building a rootfs, we need all the same depends
as do_rootfs has.
The easiest was to achieve this is to simply load them. This resolves
the failure when targeting the task from sstate.
Maciej Borzecki [Thu, 24 Jul 2014 12:27:16 +0000 (14:27 +0200)]
wic: do not overwrite autogenerated /etc/fstab with original too early
DirectImageCreator.__write_fstab() generates new /etc/fstab in sysroot
with rootfs contents. The fstab entries are generated base on the
initialn contents of /etc/fstab, plus any extra (other than / or
/boot) partitions listed in *.wks. A backup of original /etc/fstab is
done in a temp location. Subsequent call to __restore_fstab() restores
the backup copy, replacing the autogenerated one.
Calling __restore_fstab() before Wic_PartData.prepare() brings back the
original fstab before the partition image file actually is created. As
such, the autogenerated /etc/fstab will not make it to the partition.
util-linux: break out new package util-linux-findfs
We'd like to include the util-linux version of findfs in images without
having to include all of util-linux. To make this possible, break out
findfs into its own package.
util-linux: break out new package util-linux-hwclock
We'd like to include the util-linux version of hwclock in images without
having to include all of util-linux. To make this possible, break out
hwclock into its own package.
Richard Purdie [Fri, 25 Jul 2014 13:54:23 +0000 (14:54 +0100)]
bitbake: codeparser cache improvements
It turns out the codeparser cache is the bottleneck I've been observing when running
bitbake commands, particularly as it grows. There are some things we can do about
this:
* We were processing the cache with "intern()" at save time. Its actually much
more memory efficient to do this at creation time.
* Use hashable objects such as frozenset rather than set so that we can
compare objects
* De-duplicate the cache objects, link duplicates to the same object saving
memory and disk usage and improving speed
* Using custom setstate/getstate to avoid the overhead of object attribute names
in the cache file
To make this work, a global cache was needed for the list of set objects as
this was the only way I could find to get the data in at setstate object creation
time :(.
Parsing shows a modest improvement with these changes, cache load time is
significantly better, cache save time is reduced since there is now no need
to reprocess the data and cache is much smaller.
We can drop the compress_keys() code and internSet code from the shared cache
core since its no longer used and replaced by codeparser specific pieces.
Richard Purdie [Fri, 25 Jul 2014 13:50:43 +0000 (14:50 +0100)]
bitbake: cache: Don't reload the cache file since we already have this data in memory
If we're writing out merged data to disk, its safe to assume that either
we loaded the data or couldn't. Loading it again is relatively pointless
and time consuming.
Maciej Borzecki [Thu, 24 Jul 2014 12:11:50 +0000 (14:11 +0200)]
wic: squashfs partition support
It is possible to instruct wic to create a squashfs partition by setting
--fstype=squashfs in *.wks. For now this is only useable for rootfs
partitions (note that you must have squashfs support in the kernel). An
attempt to create an empty partition will produce a warning.
insane: add checking to standardize how .bbappend files do FILESEXTRAPATHS
When adding patches or config files from bbappend files, it requires
the use of FILESEXTRAPATHS, which has been an issue and failure point
for people starting to work with bitbake and oe-core.
We add checking to standardize how to use FILESEXTRAPATHS. Only the
format of:
FILESEXTRAPATHS_append := ":${THISDIR}/Your_Files_Path" or
FILESEXTRAPATHS_prepend := "${THISDIR}/Your_Files_Path:"
is acceptable.
Richard Purdie [Mon, 2 Jun 2014 20:15:01 +0000 (21:15 +0100)]
autotools: Improve configure dependency code for finding m4 files
We have an open bug about the warnings issues in builds from an sstate
cache when something like glib-2.0 gets rebuilt. The issue is that
sstate is "clever" and prunes unneeded dependencies out the tree.
For example is X depends on pkgconfig-native but we've already build X
and installed it from sstate, it will not get installed when you build Y
which depends on X.
This patch changes the logic to match the sstate behaviour and prune out
unnecessary dependencies from the scope of aclocal. This in turn removes
the warning about missing manifest files.
The issue is that this patch exposes holes in our DEPENDS in recipes,
specifically that some native tools are not listed, specifically, and
problematically, pkgconfig, gtk-doc and intltool-native in particular.
I've sent out patches against OE-Core that address the bulk of the
issues there however I'm conscious this is probably going to a bug issue
in other layers and may be too annoying to consider at this point. The
other alternative is simply to turn the warning into a debug statement.
I appreciate the code below has commented blocks, this is simply debug
I've left around for now. It will be cleaned from any final version.
Daniel BORNAZ [Thu, 24 Jul 2014 13:51:44 +0000 (15:51 +0200)]
python: fix _json module arbitrary process memory read vulnerability
http://bugs.python.org/issue21529
Python 2 and 3 are susceptible to arbitrary process memory reading by
a user or adversary due to a bug in the _json module caused by
insufficient bounds checking.
The sole prerequisites of this attack are that the attacker is able to
control or influence the two parameters of the default scanstring
function: the string to be decoded and the index.
The bug is caused by allowing the user to supply a negative index
value. The index value is then used directly as an index to an array
in the C code; internally the address of the array and its index are
added to each other in order to yield the address of the value that is
desired. However, by supplying a negative index value and adding this
to the address of the array, the processor's register value wraps
around and the calculated value will point to a position in memory
which isn't within the bounds of the supplied string, causing the
function to access other parts of the process memory.
Mozilla Netscape Portable Runtime (NSPR) before 4.10.6 allows remote
attackers to execute arbitrary code or cause a denial of service
(out-of-bounds write) via vectors involving the sprintf and console
functions.Per: http://cwe.mitre.org/data/definitions/787.html
In subdir 'gcc', Most C source files included config.h which was
generated by a rule. But no related prerequisites was added to
the C source compiling rule. There was potential building failure
while makefile enabled parallel.
The C source compiling rule used suffix rule '.c.o', but the suffix
rule doesn't support prerequisites.
https://www.gnu.org/software/make/manual/html_node/Suffix-Rules.html
We used the pattern rule '%.o : %.c' to instead, and add the config.h
as its prerequisite
We also moved the '%.o : %.c' rule down to the 'build/%.o :' rule, which
makes '%.o : %.c' rule doesn't override 'build/%.o :'.
1) bitbake perl
2) vim local.conf to tweak CC, just add redundant option.
...
CC_append = " ${HOST_CC_ARCH}"
...
3) bitbake perl
...
./miniperl -Ilib make_ext.pl lib/auto/Time/HiRes/HiRes.so MAKE=make LIBPERL_A=libperl.so LINKTYPE=dynamic
Making Time::HiRes (all)my $filename=
Deleting non-Cross makefile
Running Makefile.PL in cpan/Time-HiRes
Makefile.PL: The "xdefine" exists, skipping the configure step.
("tmp/sysroots/x86_64-linux/usr/bin/perl-native/perl5.20.0.real Makefile.PL --configure" to force the configure step)
Warning: No Makefile!
make[2]: Entering directory `tmp/work/armv5te-poky-linux-gnueabi/perl/5.20.0-r1/perl-5.20.0/cpan/Time-HiRes'
make[2]: *** No rule to make target `config'. Stop.
...
While ${CC} changed, the existance of 'xdefine' caused makefile
regeneration failed.
oeqa: Refactor test skipping decorators to use the unittest result object
In order to make the test skipping decorators independent of the oeTest object we rely on the unittest result object to construct skip, fail and error lists used by these decorators.
Created a new object getResults that analyses upper frames and retrieves the unittest result object instance, then return a list of failed, skipped and error tests.
Also removed the oetest import from decorators.py because it was no longer required.
Ross Burton [Wed, 23 Jul 2014 22:35:42 +0000 (23:35 +0100)]
debian.bbclass: inherit package
This class defines a package_name_hook implementation but as EXPORT_FUNCTIONS is
order-dependent it needs to inherit package.bbclass first to ensure that the
"base" definition in there is defined first, otherwise with a suitable inherit
order of debian and then a packaging class can result in the stub
package_name_hook being incorrectly used.
While glibc/eglibc looks for the time zone in /etc/localtime, other libc
alternatives (e.g. uclibc) may look for it in /etc/TZ. If /etc/TZ
exists, don't fall back to setting TZ to "UTC" in /etc/profile.
Ross Burton [Wed, 23 Jul 2014 22:07:22 +0000 (23:07 +0100)]
ptest-gnome: add ptest helper for GNOME packages
Many GNOME packages are using their InstalledTests pattern, where the test suite
can be trivially installed. To avoid repeating the same logic over and over,
add a class to encapsulate this.
MULTILIB_OPTIONS takes the parameters which trigger a given multilib to be
selected. It supports *one* option per multilib, '/' separated. Spaces
separate options used to generate additional multilib combinations.
Adding in all of CFLAGS to this is therefore clearly a really bad idea
but how do we fix things?
The best option I've come up with so far is a list of whitelist variables
to use to trigger the multilibs. Its populated with the standard multilibs
we support, anyone setting up an advanced multilib can populate the variable
with the correct trigger parameters.
This has the advantage of simplifying the code and allowing us to remove
the code filtering blocks since there is no longer option duplication. Testing
after this change shows a much improved sdk toolchain functionality.
Richard Purdie [Thu, 24 Jul 2014 21:10:13 +0000 (22:10 +0100)]
lib/oe/classextend: Avoid early expansion of PR values
Variables like RDEPENDS can contain EXTENDPKGV which in turn uses AUTOPR
based values. This gets set during do_package execution so we want to
defer expansion until then. The only way we can do this in the RDEPENDS
(and friends) mapping code is to subsitute a dummy value, then change it
back again. Horrible but I can't see any other way.
This resolves multilib build failures with inconsistent PR values.
Richard Purdie [Thu, 24 Jul 2014 21:09:09 +0000 (22:09 +0100)]
populate_sdk_base: Extend TOOLCHAIN_TARGET_TASK to include multilib variants
Most people expect the toolchain from a multilib build to contain multilib
components. This change makes that happen and is easy for users to override
should they want something different.
libomxil-0.9.3: Remove versioning for bellagio .so files.
The so files installed under ${libdir}/bellagio are not versioned and should
be installed without version-based symlinks so that omxregister-bellagio
can properly find and register them.
Mike Crowe [Thu, 24 Jul 2014 11:24:53 +0000 (12:24 +0100)]
allarch: Generate same package for MIPS and non-MIPS targets
LINKER_HASH_STYLE differs between MIPS and non-MIPS targets. This means
that LDFLAGS differs too. LDFLAGS is exported so it influences all task
hashes. Unfortunately this means that packages with architecture "all"
differ depending on whether they are built for a MIPS or non-MIPS target.
This causes a lot of unnecessary churn in the ipk/all directory when
switching build targets.
The simplest way to fix this is to ensure that LDFLAGS stays the same for
architecture "all" packages by clearing it. It shouldn't being used by such
packages anyway.
Ross Burton [Thu, 24 Jul 2014 19:47:08 +0000 (20:47 +0100)]
libice: fix non-deterministic libbsd dependency
libice 1.0.9 added automatic detection of arc4random(), which is in libbsd on
Linux. As this is automatic and leads to failing builds when ssstate is reused,
seed the autoconf cache as relevant to implement a PACKAGECONFIG for the
functionality.
Default to not using arc4random() as the fallback has been in use for many
years, but people interested in security may wish to turn this on to increase
the security of the X authentication cookies.
Bruce Ashfield [Tue, 22 Jul 2014 20:40:39 +0000 (16:40 -0400)]
linux-yocto/3.14: libata and generic CPU modalias handling
Updating the 3.14 yocto kernel to incorporate the following fix
and feature of interest.
5724bf17acbf x86: align x86 arch with generic CPU modalias handling 6b9a52451a78 cpu: add generic support for CPU feature based module 38367de316bb libata: support the ata host which implements a queue depth less than 32
Bruce Ashfield [Mon, 21 Jul 2014 18:09:11 +0000 (14:09 -0400)]
kernel: don't copy .so.dbg files into kernel source install
In 3.16+ x86-64 kernel builds produce a vdso64.so.dbg file. If this file is
copied into the kernel source install multiple QA failures are triggered.
Specifically, this file triggers a debug package split that results in
files installed but not shipped, and invalid .debug file errors.
By ensuring that .so files are not copied, we avoid this incorrect split
with no impact on future build phases.
Bruce Ashfield [Thu, 17 Jul 2014 19:15:32 +0000 (15:15 -0400)]
lttng-modules: update to 2.5.0
During the uprev of the yocto kernel to 3.16, lttng-modules failed to build.
To grab the latest stable content, we update to 2.5.0, and add two patches
to also make it build against 3.16+.
We also drop the older 2.3.3 lttng-modules, since it is no longer required
to support ARM builds.
Richard Purdie [Wed, 9 Jul 2014 20:26:56 +0000 (21:26 +0100)]
insane: Add build depends check
Now that we can get the task dependency tree from bitbake, we can start
to use this to strengthen our QA checks. If a dependency is added on
something which isn't in our dependency tree, that is obviously a bad
thing for example.
This patch therefore checks the RDEPENDS against the list of tasks and
ensures we do have a dependency present, if not a QA warning or error
can be issued through the usual mechanism.
The implementation is complicated by needing to resolve the RDEPENDS to
a PN using pkgdata. Its possible that can be an RPROVIDES of another
package so we need to check that too if it isn't a direct RDEPENDS.
To allow this test to work, we need to extend the do_package_qa
dependencies to include all RDEPENDS. In practise the do_package_write_*
tasks already do this so there should be no new circular dependencies or
any issues like that.
For now the issues are warnings as there are issues this finds in
OE-Core which need to be resolved and certainly will be in other layers
too. This change should simplify and assist some of Martin's dependency
scripts, the idea for this came from a discussion with Martin. It has
changed in that it doesn't just cover shlibs dependencies but checks all
dependencies.
Paul Barker [Wed, 28 May 2014 15:19:50 +0000 (15:19 +0000)]
package_ipk.bbclass: Support hierarchical feed
This patch allows for an optional new layout for ipk feed directories which I've
called a 'hierarchical feed' and is based on how Debian pools package files. It
is disabled by default and is enabled by setting IPK_HIERARCHICAL_FEED to "1".
In the traditional feed layout, package files are placed in <outdir>/<arch>/.
This can lead to several thousand files existing in a single directory which is
often a problem if developers want to upload a package feed to a shared web
hosting provider. For example, in my case, listing files via FTP only shows the
first 2000 files, breaking my scripts which attempt to upload only new and
changed files via FTP.
In the hierarchical feed, package files are written to
<outdir>/<arch>/<pkg_prefix>/<pkg_subdir>, where pkg_prefix is the first letter
of the package file name for non-lib packages or "lib" plus the 4th letter of
the package file name for lib packages (eg, 'l' for less, 'libc' for libc6).
pkg_subdir is the root of the package file name, discarding the version and
architecture parts and the common suffixes '-dbg', '-dev', '-doc', '-staticdev',
'-locale' and '-locale-*' which are listed in meta/conf/bitbake.conf.
This change relies on recent patches to opkg-utils which support hierarchical
package feeds.
Signed-off-by: Paul Barker <paul@paulbarker.me.uk> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
bitbake.conf: move BB_NUMBER_THREADS and PARALLEL_MAKE to bitbake.conf
Currently, BB_NUMBER_THREADS and PARALLEL_MAKE default to unset and
are set in local.conf. Now that we have the automatic probing,
the default values can be set in bitbake.conf and an example of
explicitly defining how many tasks to run can be moved to
local.conf.sample.extended.
[YOCTO #6217]
Signed-off-by: Roxana Ciobanu <roxana.ciobanu@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
bitbake.conf: move BB_NUMBER_THREADS and PARALLEL_MAKE to bitbake.conf
Currently, BB_NUMBER_THREADS and PARALLEL_MAKE default to unset and
are set in local.conf. Now that we have the automatic probing,
the default values can be set in bitbake.conf and an example of
explicitly defining how many tasks to run can be moved to
local.conf.sample.extended.
[YOCTO #6217]
Signed-off-by: Roxana Ciobanu <roxana.ciobanu@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Yue Tao [Tue, 22 Jul 2014 07:46:36 +0000 (15:46 +0800)]
gst-ffmpeg: Add CVE patches
Security Advisory - ffmpeg - CVE-2013-0866
The aac_decode_init function in libavcodec/aacdec.c in FFmpeg before
1.0.4 and 1.1.x before 1.1.2 allows remote attackers to have an
unspecified impact via a large number of channels in an AAC file, which
triggers an out-of-bounds array access.
The ff_add_png_paeth_prediction function in libavcodec/pngdec.c in
FFmpeg before 1.1.3 allows remote attackers to have an unspecified
impact via a crafted PNG image, related to an out-of-bounds array
access.
The ff_er_frame_end function in libavcodec/error_resilience.c in FFmpeg
before 1.0.4 and 1.1.x before 1.1.1 does not properly verify that a
frame is fully initialized, which allows remote attackers to trigger a
NULL pointer dereference via crafted picture data.
Double free vulnerability in the vp3_update_thread_context function in
libavcodec/vp3.c in FFmpeg before 0.10 allows remote attackers to have
an unspecified impact via crafted vp3 data.
The ff_h264_decode_sei function in libavcodec/h264_sei.c in FFmpeg
before 0.10 allows remote attackers to have an unspecified impact via
crafted Supplemental enhancement information (SEI) data, which triggers
an infinite loop.
The ff_combine_frame function in libavcodec/parser.c in FFmpeg before
2.1 does not properly handle certain memory-allocation errors, which
allows remote attackers to cause a denial of service (out-of-bounds
array access) or possibly have unspecified other impact via crafted
data.
The rpza_decode_stream function in libavcodec/rpza.c in FFmpeg before
2.1 does not properly maintain a pointer to pixel data, which allows
remote attackers to cause a denial of service (out-of-bounds array
access) or possibly have unspecified other impact via crafted Apple RPZA
data.
Integer overflow in the alac_decode_close function in libavcodec/alac.c
in FFmpeg before 1.1 allows remote attackers to have an unspecified
impact via a large number of samples per frame in Apple Lossless Audio
Codec (ALAC) data, which triggers an out-of-bounds array access.
Buffer overflow in FFmpeg before 0.5.6, 0.6.x before 0.6.4, 0.7.x before
0.7.8, and 0.8.x before 0.8.8 allows remote attackers to execute
arbitrary code via unspecified vectors.
The decode_init function in libavcodec/huffyuv.c in FFmpeg before 1.1
allows remote attackers to have an unspecified impact via a crafted
width in huffyuv data with the predictor set to median and the
colorspace set to YUV422P, which triggers an out-of-bounds array access.
The smacker_decode_header_tree function in libavcodec/smacker.c in
FFmpeg before 0.10 allows remote attackers to have an unspecified impact
via crafted Smacker data.
Multiple integer signedness errors in libavcodec/dsputil.c in FFmpeg
before 2.1 allow remote attackers to cause a denial of service
(out-of-bounds array access) or possibly have unspecified other impact
via crafted data.
The decode_mb function in libavcodec/error_resilience.c in FFmpeg before
0.10 allows remote attackers to have an unspecified impact via vectors
related to an uninitialized block index, which triggers an out-of-bound
write.
Array index error in the qdm2_decode_super_block function in
libavcodec/qdm2.c in FFmpeg before 1.1 allows remote attackers to have
an unspecified impact via crafted QDM2 data, which triggers an
out-of-bounds array access.
The av_probe_input_buffer function in libavformat/utils.c in FFmpeg
before 1.0.2, when running with certain -probesize values, allows remote
attackers to cause a denial of service (crash) via a crafted MP3 file,
possibly related to frame size or lack of sufficient frames to estimate
rate.
The prepare_sdp_description function in ffserver.c in FFmpeg before
1.0.2 allows remote attackers to cause a denial of service (crash) via
vectors related to the rtp format.
Signed-off-by: Yue Tao <Yue.Tao@windriver.com> Signed-off-by: Roy Li <rongqing.li@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ting Liu [Wed, 23 Jul 2014 03:58:56 +0000 (11:58 +0800)]
perf: remove CFLAGS override
definition of __SANE_USERSPACE_TYPES__ was added in kernel source by
commit e3541ec75219819d3235f80125a1a75d798ff6e1
perf tools, powerpc: Fix compile warnings in tests/attr.c
Overriding the CFLAGS makes it impossible for kernel build system to
append to it, thus making the build fail in various ways as:
| CC /.../perf/1.0-r9/perf-1.0/perf.o
| In file included from builtin.h:4:0,
| from perf.c:9:
| util/util.h:74:24: fatal error: lk/debugfs.h: No such file or directory
| #include <lk/debugfs.h>
| ^
| compilation terminated.
Richard Purdie [Wed, 23 Jul 2014 16:05:44 +0000 (17:05 +0100)]
sanity: Check for setgid/setuid TMPDIR
Building in a TMPDIR which has setgid or setuid is a bad idea. We could try and reset
the permissions but since these can also invade into other directories like the cache
or sstate, lets tell the user to fix it instead.
Ben Shelton [Wed, 23 Jul 2014 16:22:44 +0000 (11:22 -0500)]
initscripts: fix bashism in bootmisc.sh
In the commit 'initscripts: save /etc/timestamp with seconds accuracy',
a bashism was introduced in the bootmisc.sh script in the code to set
the current date from the stored value in /etc/timestamp. This causes
that operation to fail with the following message when /bin/sh is not
bash:
/etc/init.d/rc: /etc/rcS.d/S55bootmisc.sh: line 73: syntax error: bad substitution
Fixed by using pattern matching removal rather than bash-specific
substring expansion.
Signed-off-by: Richard Tollerton <rich.tollerton@ni.com> Signed-off-by: Ben Shelton <ben.shelton@ni.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
sysvinit: bootlogd: ensure /var/log/boot is created in volatiles
bootlogd does not write to /var/log/boot if it does not exist,
so if using the volatiles facility (presumed to mount /var/log under a
tmpfs), ensure that /var/log/boot gets created.
Signed-off-by: Richard Tollerton <rich.tollerton@ni.com> Signed-off-by: Ben Shelton <ben.shelton@ni.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
sysvinit: bootlogd: Don't run savelog if it's not installed
bootlogd's default log rotation code on stop requires `savelog`, which
is in debianutils, which may not be installed. If it's not installed,
don't try to perform the log rotation.
That is: in the affected code block, `savelog` is what is responsible
for creating "boot.0". When `savelog` doesn't exist, an error message
gets printed on bootup to the effect of "mv: can't find boot.0".
Signed-off-by: Richard Tollerton <rich.tollerton@ni.com> Signed-off-by: Ben Shelton <ben.shelton@ni.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The messages echoed when starting and stopping bootlogd are currently
printed regardless of the setting of VERBOSE. Adjust the initscript so
they're only printed when VERBOSE is enabled.
Signed-off-by: Richard Tollerton <rich.tollerton@ni.com> Signed-off-by: Ben Shelton <ben.shelton@ni.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Tue, 22 Jul 2014 18:38:19 +0000 (19:38 +0100)]
xcb-util-wm: upgrade to 0.4.1
Licence checksums changes as the files previously checksummed were generated and
subsequently removed from the tarball. Change the checksums to use the source
files instead.
projects / thirdparty / openembedded / openembedded-core-contrib.git / log
