PR binutils/21156: Fix illegal memory accesses in readelf when
ing a corrupt binary.
PR binutils/21156: Fix another memory access error in readelf when
parsing a corrupt binary.
Signed-off-by: Yuanjie Huang <yuanjie.huang@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Khem Raj [Tue, 23 May 2017 23:47:43 +0000 (16:47 -0700)]
xkeyboard-config: Drop explicit dep on virtual/gettext
let gettext.bbclass provide the defaults they
are good for it.
What it needs it a dep on gettext-native for intltool-native
which now intltool-native expresses correctly
Helps compile when distros disable NLS
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
When setting USE_NLS="no" and building systemd-boot
(which DEPENDS on intltool), configure fails,
complaining about missing gettext:
configure: error: GNU gettext tools not found; required for intltool
This is caused because USE_NLS="no" makes the gettext class
add gettext-minimal-native to BASEDEPENDS instead of adding
gettext-native.
Since we still would like to set USE_NLS="no" and build systemd-boot,
we add gettext-native as a dependency to intltool-native, which makes it
available for the configure script of systemd-boot fixing the problem.
tcmode-default: use SDK_SYS suffix instead of SDK_ARCH for binutils-crosssdk
Commit d2eccccb70e809d482c493922f23aef4409cfd82 has changed suffixes for
all -crosssdk packages from SDK_ARCH to SDK_SYS, but missed one line with
binutils-crosssdk. This change fixes that omission.
Signed-off-by: Denys Dmytriyenko <denys@ti.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Khem Raj [Mon, 22 May 2017 05:00:28 +0000 (22:00 -0700)]
musl: Update to latest master
* f9f686b7 have posix_spawnattr_setflags check for supported flags
* 77e895dc add no-op POSIX_SPAWN_USEVFORK to spawn.h
* 55550416 s390x: provide sigcontext struct definition
* bb439bb1 implement new posix_spawn flag POSIX_SPAWN_SETSID
* 58e2396a remove va_arg hacks in printf core with undefined behavior
* e1232f5b make ttyname[_r] return ENODEV rather than ENOENT
* 1a7fa5e5 fix regression in support for resolv.conf attempts option
* 8c44a060 fix scalbn when result is in the subnormal range
* 2577b1bc allow full-range file offsets to mmap on archs with 64-bit syscall args
* b3751c32 fix dl_iterate_phdr in static PIE binaries
* 1ca59755 fix read past end of buffer in getaddrinfo backend
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Saul Wold [Fri, 19 May 2017 23:18:46 +0000 (16:18 -0700)]
devtool/standard.py: Enable perf to be buildable
Perf is a tool build from the kernel source, which is normally available
in /work-shared/..., but when devtool is used to modify the kernel
source code, perf is not buildable since it gets an error about being unable
to add a depends to a non-exisit task do_patch.
This patch removes do_patch from the SRCTREECOVEREDTASKS and creates an empty
do_patch task to enable the VarFlags code to have someplace to attach depends
information to.
[YOCT #11120]
Signed-off-by: Saul Wold <sgw@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
This is a demo application that draws a three-dimensional spinning cube
using EGL on top of KMS, via Mesa's GBM library. A rough counterpart to
glxgears, it is a useful tool for BSP makers to try the 3D rendering
functionality based on EGL and OpenGL ES in a KMS environment, without
having to set up Wayland or X11.
It is also interesting for Yocto-based projects that do not need any
windowing environment and instead render their visual output fullscreen
to KMS directly, since they can use it as a reference.
Signed-off-by: Carlos Rafael Giani <dv@pseudoterminal.org> Signed-off-by: Ross Burton <ross.burton@intel.com>
Yeoh Ee Peng [Sun, 21 May 2017 18:22:22 +0000 (11:22 -0700)]
base.py: add assertExists and assertNotExists to oeselftest
Current osselftest print confusing assertion message when using
self.assertTrue(os.path.exists(filepath)) to test file path,
example of confusing assertion message:
AssertionError: False is not true
Add the assertExists and assertNotExists to improve assertion
message and simplify coding, using selft.assertExists(filepath),
will print meaningful assertion message:
AssertionError: <filepath> does not exist
[YOCTO #11356]
Signed-off-by: Yeoh Ee Peng <ee.peng.yeoh@intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Joshua Lock [Fri, 19 May 2017 15:53:06 +0000 (16:53 +0100)]
kernel: predefine KBUILD_BUILD_USER and KBUILD_BUILD_HOST
By exporting KBUILD_BUILD_USER with a pre-defined value we improve the
reproducibility of the kernel and remove the requirement for whoami in the
HOSTTOOLS.
KBUILD_BUILD_HOST also helps improve the reproducibility of the kernel.
For more kernel reproducibility options see:
https://lwn.net/Articles/437864/
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Paul Eggleton [Fri, 19 May 2017 04:11:48 +0000 (16:11 +1200)]
scriptutils: fix fetch_uri() to work with RSS
Since recipe-specific sysroots were implemented, devtool add and devtool
upgrade operations that fetch from a URL that requires native sysroot
dependencies will fail to work as there is no recipe-specific sysroot
set up for them during fetching. An example was any URL pointing to a
tarball compressed with xz, e.g. devtool upgrade on gnutls.
The most expedient way to fix this is to set up a dummy recipe-specific
sysroot to use for the fetch/unpack operations. We do this in the same
manner as bitbake -b does, so we're just taking all of the sysroot
components available and creating a sysroot from those rather than
ensuring the correct dependencies are there - this means that we're
still going to have problems if e.g. xz-native hasn't been built yet,
but that issue will be trickier to solve and is tracked separately.
Fixes [YOCTO #11474].
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Max Krummenacher [Wed, 17 May 2017 20:11:34 +0000 (22:11 +0200)]
u-boot-mkimage: fix nativesdk build
If building for nativesdk the wrong rss sysroot is used leading the
following error message.
| ERROR: oe_runmake failed
| In file included from tools/imximage.c:13:0:
| include/image.h:1024:27: fatal error: openssl/evp.h: No such file or directory
| # include <openssl/evp.h>
Tools needed on the build host (script/basic/fixdep) and code compiled
for the SDK machine are both built with the build host's compiler,
leading to additinal errors.
Adding CROSS_COMPILE="${HOST_PREFIX}" and using the cross-compiler for
the SDK_ARCH fixes the build error.
The resulting binary in the SDK is working.
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Ed Bartosh [Wed, 17 May 2017 09:07:47 +0000 (12:07 +0300)]
systemd-boot: create output dir if it doesn't exist
build_efi_cfg function creates configuration files for
systemd-boot entries in 'S' directory. This directory
may not exist when api is called, which breaks the build.
Creating the directory if it doesn't exist should fix
this issue.
Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Markus Lehtonen [Mon, 15 May 2017 11:18:45 +0000 (14:18 +0300)]
scripts/buildstats-diff: support optimized rusage values
Buildstats from oe-build-perf-test results have been optimized to not
have child rusage values at all. There, rusage is the sum of parent and
child rusage values. This patch makes buildstats-diff compatible with
this format.
[YOCTO #11355]
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Markus Lehtonen [Mon, 15 May 2017 11:18:44 +0000 (14:18 +0300)]
oe-build-perf-report: implement --dump-buildstats
For dumping buildstats from the test runs being reported. The output
directory where buildstats are copied is 'oe-build-perf-buildstats/'.
Buildstats can be then further analyzed with buildstats-diff script, for
example.
[YOCTO #11355]
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Markus Lehtonen [Mon, 15 May 2017 11:18:41 +0000 (14:18 +0300)]
oeqa.utils.git: use --verify in rev_parse()
We use rev-parse for turning git object names into SHA-1 and checking
their existence. Using --verify option makes sure git-rev-parse does
what we expect.
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Paul Eggleton [Sun, 14 May 2017 21:59:53 +0000 (09:59 +1200)]
recipetool: create: extract name of package from a repository
For git repositories in the absence of any other indicator, it's not an
unreasonable assumption that the name of the repository is the name of
the software package it contains, so use that as PN if we don't have
anything else.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Paul Eggleton [Sun, 14 May 2017 21:59:52 +0000 (09:59 +1200)]
recipetool: create: skip values extracted from spec files containing macros
If a value we extract from a spec file contains an unexpanded macro
(e.g. %{macroname}) then we should discard it since we're not seeing the
actual value and we don't have an easy way of expanding it at the
moment.
This fixes for example getting %{name} as the recipe name when running
the following:
Khem Raj [Sat, 6 May 2017 17:34:14 +0000 (17:34 +0000)]
glew: Compile with -D_GNU_SOURCE
Required with glibc 2.26
when x11 is not in distro features it errors
out due to mising CLOCK_REALTIME
error: 'CLOCK_REALTIME' undeclared (first use in this function);
did you mean '_XOPEN_REALTIME'?
if (clock_gettime(CLOCK_REALTIME, &ts) == -1)
^~~~~~~~~~~~~~
_XOPEN_REALTIME
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Jackie Huang [Wed, 10 May 2017 06:29:11 +0000 (14:29 +0800)]
icu: Use LARGE_BUFFER_MAX_SIZE for cmd
The previous patch used LARGE_BUFFER_MAX_SIZE instead
of SMALL_BUFFER_MAX_SIZE for cmd in function pkg_installLibrary,
which only fixed some of the cases when the command line
is too long, some other cases indicate that the
LARGE_BUFFER_MAX_SIZE is also needed in pkg_installCommonMode
and pkg_installFileMode to avoid overflow:
Maxin B. John [Mon, 8 May 2017 11:46:57 +0000 (14:46 +0300)]
connman: upgrade to 1.34
1.33 -> 1.34
1. Refreshed 0001-Fix-compile-on-musl-with-kernel-4.9-headers.patch
2. Removed upstreamed patch:
0003-stats-Fix-bad-file-descriptor-initialisation.patch
3. Provided PACKAGECONFIGs for nftables and iptables support
4. Add new patch to fix build with nftables:
0001-firewall-nftables-fix-build-with-libnftnl-1.0.7.patch
Signed-off-by: Maxin B. John <maxin.john@intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Marek Vasut [Fri, 5 May 2017 17:35:48 +0000 (19:35 +0200)]
automake: Backport perl 5.22 fix
Backport 13f00eb4493c "automake: port to Perl 5.22 and later"
from automake upstream to fix build with perl 5.22 .
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Ross Burton <ross.burton@intel.com> Cc: Richard Purdie <richard.purdie@linuxfoundation.org> Signed-off-by: Ross Burton <ross.burton@intel.com>
Cody P Schafer [Thu, 4 May 2017 17:50:42 +0000 (13:50 -0400)]
cmake.bbclass: use `cmake --build` to build & install
Rather than presuming `make` is the generator, use cmake's generic
`cmake --build` feature (which knows to call the appropriate generator).
Both DESTDIR and VERBOSE still behave as intended when used as
environment variables instead of make variable-arguments.
As cmake-based builds don't do any configuration with `make`
invocations, we only pass `PARALLEL_MAKE{,INST}` (via a
EXTRA_OECMAKE_BUILD variable) to the underlying build tool. Make &
ninja support the same `-j N` option (and a few others), so this does
happen to work for both.
This makes it more straight forward for others to select other cmake
generators (many folks have been reaching for `ninja` lately).
CC: Andre McCurdy <armccurdy@gmail.com> Signed-off-by: Cody P Schafer <dev@codyps.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Add missing parameters -c to and remove non-existent
parameter -i from usage string for 'wic create'.
Signed-off-by: Andreas Reichel <andreas.reichel.ext@siemens.com> Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com> Signed-off-by: Daniel Wagner <daniel.wagner@siemens.com> Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
* optparse is deprecated and will not be developed further
(see: https://docs.python.org/2/library/optparse.html)
* argparse supports subcommands, which simplifies definition of
arguments and options
* reimplement help mechanism through sub-subcommands
[YOCTO #9636]
Signed-off-by: Andreas Reichel <andreas.reichel.ext@siemens.com> Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com> Signed-off-by: Daniel Wagner <daniel.wagner@siemens.com> Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
The following warning occurs when building with meta-zephyr
with MACHINE set to arduino-101-sss:
WARNING: /srv/sdc/builds/11319/meta/recipes-extended/ghostscript/ghostscript_9.20.bb: Unable to get checksum for ghostscript SRC_URI entry objarch.h: file could not be found
This is due to the the TARGET_ARCH = "arc" for meta-zephyr is not
supported by ghostscript and causing bitbake unable to locate the
correct config file during recipe parse.
Adding checker in the recipe to raise an exception if the target
architecture is "arc". This would then only display an error if
someone specifically tries to build the recipe:
ERROR: ghostscript was skipped: incompatible with host arc-yocto-elf (not in COMPATIBLE_HOST)
[YOCTO #11344]
Signed-off-by: Chang Rebecca Swee Fun <rebecca.swee.fun.chang@intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Huang Qiyu [Tue, 9 May 2017 02:22:25 +0000 (10:22 +0800)]
libpciaccess: 0.13.4 -> 0.13.5
1) Upgrade libpciaccess from 0.13.4 to 0.13.5.
3) Delete three patches, since they are integrated upstream.
0001-Include-config.h-before-anything-else-in-.c.patch
0002-Fix-quoting-issue.patch
0003-linux_sysfs.c-Include-limits.h-for-PATH_MAX.patch
Signed-off-by: Huang Qiyu <huangqy.fnst@cn.fujitsu.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Catalin Enache [Mon, 8 May 2017 13:42:59 +0000 (16:42 +0300)]
ghostscript: CVE-2016-8602, CVE-2017-7975
The .sethalftone5 function in psi/zht2.c in Ghostscript before 9.21 allows remote
attackers to cause a denial of service (application crash) or possibly execute
arbitrary code via a crafted Postscript document that calls .sethalftone5 with an
empty operand stack.
Artifex jbig2dec 0.13, as used in Ghostscript, allows out-of-bounds writes because
of an integer overflow in the jbig2_build_huffman_table function in jbig2_huffman.c
during operations on a crafted JBIG2 file, leading to a denial of service (application
crash) or possibly execution of arbitrary code.
1) Upgrade libgpg-error from 1.26 to 1.27.
2) One file's LIC_FILES_CHKSUM checksums changed (src/gpg-error.h.in),but LICENSE remains the same.
3) Delete PR assignmen,since the version is upgraded.
Signed-off-by: Huang Qiyu <huangqy.fnst@cn.fujitsu.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Prior versions of python do not support openssl 1.1; updating to
Python 3.6 on the other hand is a lot more involved, and so should
be done by a specialist/maintainer.
LICENSE checksum change due to copyright years.
Drop upstreamed python3-fix-CVE-2016-1000110.patch
Aníbal Limón [Mon, 1 May 2017 17:11:23 +0000 (12:11 -0500)]
selftest/devtool: change write to append config in virtual_kernel_modify
When uses write it creates a full new configuration in selftest.inc
causing to fail populate_sdk_ext linux-yocto because the signature
of the sstate changes.
[YOCTO #11300]
Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Patrick Ohly [Wed, 3 May 2017 13:38:53 +0000 (15:38 +0200)]
QemuRunner: avoid tainting os.environ
That a utility function permanently changes the process environment is
bad style and leads to subtle, hard to debug problems.
For example, we had one oe-selftest which used runqemu() with an
override for DEPLOY_DIR_IMAGE. Another test then just called runCmd()
and ended up passing the wrong DEPLOY_DIR_IMAGE set earlier in
os.environ.
The approach used here is to pass the desired environment dict to the
launch() method as a new, optional parameter, which then gets passed
on to subproject.Popen(). The modified env variables do not get
logged, as before.
[YOCTO #11443]
Signed-off-by: Patrick Ohly <patrick.ohly@intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Chen Qi [Tue, 9 May 2017 09:31:36 +0000 (17:31 +0800)]
cve-check.bbclass: make warning contain CVE IDs
When warning users about unpatched CVE, we'd better put CVE IDs into
the warning message, so that it would be more straight forward for the
user to know which CVEs are not patched.
So instead of:
WARNING: gnutls-3.5.9-r0 do_cve_check: Found unpatched CVE, for more information check /path/to/workdir/cve/cve.log.
We should have:
WARNING: gnutls-3.5.9-r0 do_cve_check: Found unpatched CVE (CVE-2017-7869), for more information check /path/to/workdir/cve/cve.log.
Signed-off-by: Chen Qi <Qi.Chen@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Patrick Ohly [Fri, 5 May 2017 10:25:25 +0000 (12:25 +0200)]
archiver.bbclass: do not cause kernel rebuilds
Adding or removing archiver.bbclass from a build configuration causes
rebuilds of linux-yocto-based kernels because of the
do_kernel_configme->do_unpack_and_patch task dependency.
This particular dependency can be ignored for the do_kernel_configme
sstate signature calculcation. Idea for the fix from Richard Purdie.
Note that building the kernel and adding archiver.bbclass later to
archive sources leads to do_unpack_and_patch running after
do_kernel_configme (because that already ran in the first build),
which might be problematic. This is independent of the change here.
The use case in YOCTO #11441 is to removed archiver.bbclass between a
production build with archiving enabled and builds via oe-selftests
without archiving. That direction is fine.
Fixes: YOCTO #11441 Signed-off-by: Patrick Ohly <patrick.ohly@intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Patrick Ohly [Fri, 5 May 2017 10:25:24 +0000 (12:25 +0200)]
archiver.bbclass: various fixes for original+diff mode
The diff.gz gets created in do_unpack_and_patch, but
do_deploy_archives did not depend on it, so there was a race
condition. For example, "bitbake linux-intel:do_deploy_archives"
without a prior "bitbake linux-intel:do_kernel_configme" did not
deploy the diff.gz.
When do_unpack_and_patch ran first, it failed because the output
directory didn't exist yet and the error was not detected because the
result of the diff command wasn't checked.
Changing the current working directory in create_diff_gz() without
returning to the original directory caused warnings like this:
WARNING: linux-intel-... do_unpack_and_patch: Task do_unpack_and_patch changed cwd to .../tmp-glibc/work-shared/intel-corei7-64
Signed-off-by: Patrick Ohly <patrick.ohly@intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Previously, do_ar_recipe ran again unnecessarily when adding or
removing classes like buildhistory.bbclass, because that changes the
BBINCLUDED variable which do_ar_recipe uses to find .bbappend files.
This is both extra work and also sometimes triggered "basehash
changed" errors (seen under oe-selftest, which adds machine.inc and
bblayers.inc) because BBINCLUDED is special and does not cause
the basehash to be recalculated.
The file *content* already was not considered in the task signature,
instead relying indirectly on PF (which includes the revision assigned
by a PR server) to ensure that a new versioned source archive gets
created each time there is a rebuild.
Therefore it makes sense to use the same mechanism and also ignore the
file *list*, i.e. exclude BBINCLUDED from the task signature.
Signed-off-by: Patrick Ohly <patrick.ohly@intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Ross Burton [Tue, 16 May 2017 15:27:00 +0000 (16:27 +0100)]
opkg-utils: don't generate PN-dev and PN-staticdev
opkg-utils ends up in any opkg/rpm image with package management as it's the
provider for update-alternatives. If dev-pkgs is enabled then opkg-utils-dev
will get installed, which is empty but will subsequently pull python3-dev into
the image (as opkg-utils-python depends on python3).
As this can result in all of Python appearing in otherwise small images, don't
generate these pointless packages.
Signed-off-by: Ross Burton <ross.burton@intel.com>
