Richard Purdie [Sat, 11 Aug 2018 10:30:28 +0000 (11:30 +0100)]
glibc: Add make-native depends
glibc needs make >= 4 yet some of our build workers have older versions of
make. Add a make-native dependency to work around this until all our supported
distros have a recent version of make.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Martin Jansa [Thu, 9 Aug 2018 12:16:59 +0000 (12:16 +0000)]
dropbear.inc: add dependency on virtual/crypt to fix build with glibc-2.28
configure tests crypt() existence with:
dnl We test for crypt() specially. On Linux (and others?) it resides in libcrypt
dnl but we don't want link all binaries to -lcrypt, just dropbear server.
dnl OS X doesn't need -lcrypt
AC_CHECK_FUNC(crypt, found_crypt_func=here)
AC_CHECK_LIB(crypt, crypt,
[
CRYPTLIB="-lcrypt"
found_crypt_func=here
])
AC_SUBST(CRYPTLIB)
if test "t$found_crypt_func" = there; then
AC_DEFINE(HAVE_CRYPT, 1, [crypt() function])
fi
but that silently fails with glibc-2.28 and a bit later do_compile fails with;
http://errors.yoctoproject.org/Errors/Details/185895/
Martin Jansa [Thu, 9 Aug 2018 11:44:47 +0000 (11:44 +0000)]
glide: add INSANE_SKIP for textrel
* I'm not using glide, so I'm not going to fix it proplerly,
it was just bothering me in world builds
* this is reproducible only with ptest in DISTRO_FEATUREs (for aarch64
issue) and included security_flags.inc, more specifically with the PIE
flags, so alternative work around is:
SECURITY_CFLAGS_pn-glide = "${SECURITY_NOPIE_CFLAGS}"
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Martin Jansa [Thu, 9 Aug 2018 11:44:46 +0000 (11:44 +0000)]
go(-dep): add INSANE_SKIP for textrel
* I'm not using go or go-dep, so I'm not going to fix it proplerly,
it was just bothering me in world builds
* this is reproducible only with ptest in DISTRO_FEATUREs (for aarch64
issue) and included security_flags.inc, more specifically with the PIE
flags, so alternative work around is:
SECURITY_CFLAGS_pn-go = "${SECURITY_NOPIE_CFLAGS}"
SECURITY_CFLAGS_pn-go-dep = "${SECURITY_NOPIE_CFLAGS}"
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Mon, 13 Aug 2018 17:20:54 +0000 (18:20 +0100)]
classes: sanity-check LIC_FILES_CHKSUM
We assume that LIC_FILES_CHKSUM is a file: URI but don't actually verify this,
which can lead to problems if you have a URI that resolves to a path of / as
Bitbake will then dutifully checksum / recursively.
[ YOCTO #12883 ]
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Mon, 13 Aug 2018 23:59:39 +0000 (00:59 +0100)]
bzip2: use Yocto Project mirror for SRC_URI
The bzip.org domain expired and is now a holding site for adverts, so we can't
trust a tarball that appears on that site (luckily we have source checksums to
detect this).
For now, point SRC_URI at the tarball in the Yocto Project source mirror, but
set HOMEPAGE and UPSTREAM_CHECK_URI to the sourceware.org/bzip2/ page which
apparently will be resurrected as the new canonical home page.
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Mikko Rapeli [Mon, 13 Aug 2018 09:42:17 +0000 (12:42 +0300)]
perf: fail if src path does not exist
A missing src directory from a broken kernel recipe resulted
only in a warning:
WARNING: copyfile: stat of /home/builder/src/tmp-glibc/work-shared/target/kernel-source/tools/arch failed ([Errno 2] No such file or directory: '/home/builder/src/tmp-glibc/work-shared/target/kernel-source/tools/arch')
and the build failed horribly.
With this change it's an error which can not be missed:
ERROR: perf-1.0-r9 do_configure: Path does not exist: /home/builder/src/tmp-glibc/work-shared/target/kernel-source/tools/arch. Maybe PERF_SRC does not match the kernel version.
ERROR: perf-1.0-r9 do_configure: Function failed: copy_perf_source_from_kernel
ERROR: Logfile of failure stored in: /home/builder/src/tmp-glibc/work/target-linux/perf/1.0-r9/temp/log.do_configure.21083
NOTE: recipe perf-1.0-r9: task do_configure: Failed
ERROR: Task (/home/builder/src/poky/meta/recipes-kernel/perf/perf.bb:do_configure) failed with exit code '1'
To get get perf compiling from a custom kernel, a perf.bbappend can be
created which defines PERF_SRC as list of files and directories needed
from kernel source tree to compile perf. This varies between kernel
versions.
Signed-off-by: Mikko Rapeli <mikko.rapeli@bmw.de> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Chen Qi [Mon, 13 Aug 2018 08:54:37 +0000 (16:54 +0800)]
parselogs.py: output correct log location
The log entry in results is altered to remove 'target_logs'. This
causes wrong log location in output.
e.g.
AssertionError: 1 != 0 : Log: /path/to/image/1.0-r0/postinstall.log
But when user wants to check the log, the user will find the log is
not present. The actual log file is /path/to/image/1.0-r0/target_logs/postinstall.log.
So fix to use the correct log location.
Signed-off-by: Chen Qi <Qi.Chen@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Chen Qi [Mon, 13 Aug 2018 09:43:23 +0000 (17:43 +0800)]
man-db: rdepend on base-passwd to ensure installation order
Make man-db rdepend on base-passwd to ensure that base-passwd is
installed before man-db at rootfs time. This is to avoid the following
warning at rootfs time.
warning: user man does not exist - using root
Signed-off-by: Chen Qi <Qi.Chen@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Error: Transaction check error:
file /usr/share/pkgconfig/mobile-broadband-provider-info.pc from install of \
lib32-mobile-broadband-provider-info-dev.core2_32 conflicts with file from package \
mobile-broadband-provider-info-dev.core2_64
Signed-off-by: Zhixiong Chi <zhixiong.chi@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Zhixiong Chi [Mon, 13 Aug 2018 07:04:23 +0000 (00:04 -0700)]
libevent: fix the multilib header conflict
Error: Transaction check error:
file /usr/include/event2/event-config.h conflicts between attempted installs of libevent-dev-2.1.8-r0.skylake_64 and lib32-libevent-dev-2.1.8-r0.x86
The conflict is the size macro definition between 32bit and 64bit
such as:
< #define EVENT__SIZEOF_LONG 8
> #define EVENT__SIZEOF_LONG 4
< #define EVENT__SIZEOF_PTHREAD_T 8
> #define EVENT__SIZEOF_PTHREAD_T 4
Signed-off-by: Zhixiong Chi <zhixiong.chi@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Chen Qi [Wed, 1 Aug 2018 05:25:35 +0000 (13:25 +0800)]
base-files: fix handling of resize
The current handling of resize is incorrect. Using `resize > /dev/null
2>&1 && resize > /dev/null' will cause the second resize command to not
execute because 'resize > /dev/null 2>&1' will fail for resize utility
from busybox.
What we really should do is just to check whether ${bindir}/resize
is executable and execute it if so. Using '-x' is sufficient.
Signed-off-by: Chen Qi <Qi.Chen@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Andrej Valek [Thu, 9 Aug 2018 08:06:37 +0000 (10:06 +0200)]
libxml2: Fix CVE-2018-14404
Fix nullptr deref with XPath logic ops
If the XPath stack is corrupted, for example by a misbehaving extension
function, the "and" and "or" XPath operators could dereference NULL
pointers. Check that the XPath stack isn't empty and optimize the
logic operators slightly.
CVE: CVE-2018-14404 Signed-off-by: Andrej Valek <andrej.valek@siemens.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
If multilib scripts handle more than one file per package, the variable
ALTERNATIVE_${PN} will be overwritten and there will be only one symbol
link file. Append to the variable to avoid this.
Signed-off-by: Zhixiong Chi <zhixiong.chi@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Jaewon Lee [Mon, 30 Jul 2018 21:21:53 +0000 (14:21 -0700)]
kernel-yocto.bbclass: Adds oe-local-files path (devtool) to include directives
The devtool-source class moves all local files specified in SRC_URI to
an oe-local-files directory. When using devtool and a recipe space kernel-meta,
devtool modify throws an error because the paths the kernel-yocto class
is looking for feature directories in, don't include the oe-local-files
directory which devtool is using.
This patch checks for feature directories in oe-local-files,
and if present, adds that path to include directives.
[YOCTO #12855]
Signed-off-by: Jaewon Lee <jaewon.lee@xilinx.com> Signed-off-by: Alejandro Enedino Hernandez Samaniego <alejandr@xilinx.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Jaewon Lee [Thu, 9 Aug 2018 23:41:29 +0000 (16:41 -0700)]
devtool-source.bbclass: Support kernel-fragments/patch not in SRC_URI
When using a recipe space kernel-meta, scc files are added through
SRC_URI, but they may include corresponding kernel fragments or patches
that are not necessarily in SRC_URI.
For bitbake, this is not a problem because the kernel-yocto class adds
the path where the .scc file was found to includes which consequentially
makes the .cfg, .patch file available to the kernel build.
However, when using devtool, only files specified in SRC_URI are copied
to oe-local-files in devtool's workspace. So if the cfg/patch file is not in
SRC_URI, it won't be copied, causing a kernel build failure when trying
to find it.
This fix parses local .scc files in SRC_URI, copies the corresponding
.cfg/.patch file to devtool's workdir, and also adds it to local_files
so it is available when doing a devtool build for the kernel.
[YOCTO #12858]
v2: also supporting patch not in SRC_URI
v3: fix spacing issues
Signed-off-by: Jaewon Lee <jaewon.lee@xilinx.com> Signed-off-by: Alejandro Enedino Hernandez Samaniego <alejandr@xilinx.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
then it fails due to being unable to find ccache. The references to ccache are
coded into libtool-cross but the sstate checksum doesn't reflect this due to the
way the class is coded (output should be the same regardless).
The simplest solution is to remove references to ccache from the libtool script.
The output then works regardless of whether ccache is present or not. The
libtool-cross script is only used in a handful of cases (most of the time its
dynamically generated by autoconf) so any performance issue is minor.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 9 Aug 2018 16:58:07 +0000 (16:58 +0000)]
sstate: Remove DEPLOY_DIR_IMAGE from SSTATE_DUPWHITELIST
Replace the generic whitelist entry with entries for the three specific
'problem' cases in OE-Core. This means the general DEPLOY_DIR_IMAGE entry
doesn't mask problems for others as was recently encoutered by users
reported on irc. In the whitelisted cases they occur only in multilib builds
and the files are identical.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 9 Aug 2018 12:17:39 +0000 (12:17 +0000)]
oeqa/esdk/devtool: Drop OETestDepends usage
OETestDepends doesn't work with parallelism and in this case we don't
really need this dependency, it would just short out some tests quickly
in the rare case the esdk environment was broken.
Currently this is masking tests which is a much worse problem and we
can't make OETestDepends work reliably with parallelism so drop the
dependencies.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Urs Fässler [Wed, 8 Aug 2018 07:33:52 +0000 (09:33 +0200)]
cmake: fix compiling some C++ projects with Yocto SDK and GCC
Setting CMAKE_SYSROOT in the toolchain file allows CMake to correctly
remove user-provided system include directories pointing to
<sysroot>/usr/include. The mentioned projects failed with "stdlib.h:
No such file or directory #include_next <stdlib.h>".
Andre McCurdy [Wed, 8 Aug 2018 18:15:12 +0000 (11:15 -0700)]
openssl_1.0: drop unnecessary call to perlpath.pl from do_configure()
The perlpath.pl script is used to patch the #! lines in all perl
scripts in the utils directory. However, as these scripts are run via
e.g. "perl foo.pl", they don't actually rely on the #! path to be
correct (which can be confirmed by the observation that the path is
currently being set to ${STAGING_BINDIR_NATIVE}/perl, which doesn't
exist).
Signed-off-by: Andre McCurdy <armccurdy@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
SERIAL_CONSOLE was already deprecated in 2013, yet still some
machine configuration files were using it. This patch replaces
it with SERIAL_CONSOLES, which is the successor.
The default value in systemd-serialgetty.bb can also be safely
transitioned from SERIAL_CONSOLE to SERIAL_CONSOLES, as this
recipe already uses SERIAL_CONSOLES within do_install().
The documentation seems to be already up do date.
beaglebone-yocto.conf in the bsp-guide already uses
SERIAL_CONSOLES. The ref-manual redirects from SERIAL_CONSOLE
to SERIAL_CONSOLES.
[YOCTO #12653]
Signed-off-by: Maciej Pijanowski <maciej.pijanowski@3mdeb.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Daniel Díaz [Thu, 9 Aug 2018 17:14:54 +0000 (12:14 -0500)]
multilib_header: recognize BPF as a target
When building with `clang -target bpf` using the
multilib_header, a recursion was unavoidable because
bits/wordsize.h would #include itself, still lacking
a definition for __MHWORDSIZE or __WORDSIZE.
Signed-off-by: Daniel Díaz <daniel.diaz@linaro.org> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Daniel Díaz [Thu, 9 Aug 2018 17:14:53 +0000 (12:14 -0500)]
glibc: Make bits/wordsize.h multilibbed again
As reported by ChenQi, leaving bits/wordsize.h out of being
multilibbed introduced a problem in building the SDK for
arm64:
Error: Transaction check error:
file /usr/include/bits/wordsize.h conflicts between attempted installs of lib32-libc6-dev-2.27-r0.armv7vet2hf_vfp and libc6-dev-2.27-r0.aarch64
Martin Jansa [Thu, 9 Aug 2018 13:39:45 +0000 (13:39 +0000)]
man-pages: respect api-documentation
* let manpages.bbclass to enable manpages PACKAGECONFIG based on
api-documentation DISTRO_FEATURES
PACKAGECONFIG_append_class-target = " ${@bb.utils.contains('DISTRO_FEATURES', 'api-documentation', 'manpages','', d)}"
* it's true that building man-pages without manpages being enabled
doesn't make much sense, but it's included through couple
packagegroups:
meta/recipes-core/packagegroups/packagegroup-self-hosted.bb: man-pages \
meta/recipes-extended/packagegroups/packagegroup-core-lsb.bb: man-pages \
or in world even for people who might not be interested
in man-pages
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Khem Raj [Wed, 8 Aug 2018 22:49:13 +0000 (15:49 -0700)]
lib/oe: Fix collections ABCs DeprecationWarning in Python 3.7+
- Prefer collections.abc (new in Python 3.3) over collections for abstract base classes
- In Python 3.8, the abstract base classes in collections.abc will no longer be exposed in
the regular collections module. This will help create a clearer distinction between
the concrete classes and the abstract base classes."
foocampo [Sat, 4 Aug 2018 00:47:40 +0000 (19:47 -0500)]
package: skip strip on signed kernel modules
Executing strip action on kernel modules removes the signature.
Is not possible to strip and keep the signature, therefore avoid
strip signed kernel modules.
Signed-off-by: Omar Ocampo <omar.ocampo.coronado@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Tue, 7 Aug 2018 14:26:20 +0000 (14:26 +0000)]
sstate: Add sstate usage summary to the build
Currently the user has no indication of how much sstate was already present
or that would be used by the build. This change adds some summary information
so that the user can see how much reuse is occurring. To fully work it
needs some extra information from a recent bitbake commit but this is
optional.
When combined with bitbake --dry-run this feature can be used to check
if sstate would be reused in a build.
[YOCTO #12749]
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Andre McCurdy [Tue, 31 Jul 2018 01:28:07 +0000 (18:28 -0700)]
openssl_1.0: drop unnecessary dependency on makedepend-native
The openssl Configure script will only select standalone makedepend
(vs running "$CC -M") when building with gcc < 3.x or with an Apple
Xcode version which predates the switch to clang (in approx 2010?).
Neither of these cases are possible when building under OE, therefore
the dependency on makedepend-native can be dropped (ie align the
openssl 1.0 recipe with the 1.1 recipe, which has dropped the
makedepend-native dependency already).
Signed-off-by: Andre McCurdy <armccurdy@gmail.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Khem Raj [Fri, 3 Aug 2018 17:30:14 +0000 (10:30 -0700)]
libidn2: Fix charset.alias issue with musl
Fixes
ERROR: libidn2-2.0.5-r0 do_package: QA Issue: libidn2: Files/directories were installed but not shipped in any packag
e:
/usr/lib/charset.alias
Please set FILES such that these items are packaged. Alternatively if they are unneeded, avoid installing them or del
ete them within do_install.
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Paulo Neves [Mon, 6 Aug 2018 11:26:58 +0000 (13:26 +0200)]
context.py: Do not mask exceptions needlessly.
There were a lot of assumptions in the controller import
code of context.py which were not true anymore. These
assumptions reflected themselves by catching exceptions and
re-raising them with unhelpful error messages in other parts
of the code.
This commit does not fix the classes controller classes that
became broken after the refactor but at least it allows for
the exceptions to be thrown where the imports fail, thus
actually showing what exactly went wrong with the import.
An example of such an improvement is that before if the
controller class failed during it's init contructor the
controller would just be skipped and the task would just
complain it could not find the controller. Now for example,
if there is a NamerError due to a variable not being
declared, the user will get that report.
Signed-off-by: Paulo Neves <ptsneves@gmail.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Chen Qi [Mon, 6 Aug 2018 03:17:25 +0000 (11:17 +0800)]
oe_syslog.py: fix for syslog-ng
When using syslog-ng as the syslog provider, oe_syslog test case fails
because it cannot find the syslog daemon. This is because it greps for
'syslogd' but syslog-ng's daemon is 'syslog-ng'. So fix it to check both
'syslogd' and 'syslog-ng'.
Also, when the test case fails, what I get is:
| AssertionError: 1 != 0 : No syslogd process; ps output:
<empty here>
This does not help user. The output is actually from the 'PS | GREP' command.
And when the 'PS | GREP' command fails, the output is always empty. So also fix
this problem. After the change, it looks like:
| AssertionError: False is not true : No syslog daemon process; ps output:
| PID USER VSZ STAT COMMAND
| 1 root 16476 S {systemd} /sbin/init
| 2 root 0 SW [kthreadd]
| 3 root 0 IW [kworker/0:0]
...
Signed-off-by: Chen Qi <Qi.Chen@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Using os.path.getmtime() will dereference symbolic links in an attempt
to get the last modified time. This can cause errors if the target
doesn't exist, or worse map to some absolute build host path which would
make a build not reproducible.
Signed-off-by: Joshua Watt <JPEWhacker@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Nathan Rossi [Thu, 2 Aug 2018 08:55:04 +0000 (18:55 +1000)]
devicetree.bbclass: User/BSP device tree source compilation class
This bbclass implements the device tree compilation for user provided
device trees. In order to use this class, it should be inherited in a
BSP recipe which provides the sources. The default setup enables
inclusion of kernel device tree sources (though can be disabled by the
recipe by overriding DT_INCLUDE or KERNEL_INCLUDE).
This provides an additional mechanism for BSPs to provide device trees
and device tree overlays for their target machines. Whilst still
enabling access to the kernel device trees for base SoC includes and
headers.
This approach to providing device trees has benefits for certain use
cases over patching the device trees into the kernel source.
* device trees are separated from kernel source, allows for selection of
kernel and or kernel versions without needing to explicitly patch the
kernel (or appending to the kernel recipes).
* providing device trees from separate sources, from the layer,
generated by the recipe or other recipes.
This class also implements some additional features that are not
available in the kernel-devicetree flow. This includes population of
device tree blobs into the sysroot which allows for other recipes to
consume built dtbs (e.g. U-Boot with EXT_DTB compilation), device tree
overlay compilation and customizing DTC compilation args (boot
cpu/padding/etc.).
Signed-off-by: Nathan Rossi <nathan@nathanrossi.com> Acked-by: Martin Hundebøll <mnhu@prevas.dk> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Tue, 7 Aug 2018 11:01:36 +0000 (11:01 +0000)]
selftest/package: Fix hardlink test when using sstate
The build target won't cause the package task to run if things were already
built and in sstate. Ensure we run the package task explicitly to ensure
the test works as intended.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This causes regression on build machines where libunistring is installed
on host. It is also because gnuts is using non standard AC macro called
AC_LIB_HAVE_LINKFLAGS to detect this library and it confusing cross builds.
Alex Kiernan [Thu, 2 Aug 2018 14:42:29 +0000 (14:42 +0000)]
net-tools: Ensure quilt only operates on local patches
When net-tools is built and the local patches haven't been applied, then
`quilt pop -a` will climb parent directories attempting to find a
{patches/.pc} directory; if this succeeds then we end up popping off
some completely unrelated patches from a parent directory.
Ensure that we do have a local patches directory and then turn off the
directory climbing so that we avoid this problem.
Signed-off-by: Alex Kiernan <alex.kiernan@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Joshua Watt [Fri, 3 Aug 2018 14:10:30 +0000 (09:10 -0500)]
alsa-lib: Cleanup packaging
Cleans up the packaging by moving libasound.so.2 back into the alsa-lib
package which was previously empty.
Previously, it was difficult to create an image that had libasound.so.2,
then create an SDK from that image that had the proper development
files, because the only way to get libasound.so.2 was to do:
IMAGE_INSTALL += "libasound"
This however caused a problem because all of the development files that
would be desired in the SDK were located in alsa-lib-dev, which wouldn't
be included because alsa-lib wasn't included, and it was impossible to
include alsa-lib because it was an empty package that was culled.
Signed-off-by: Joshua Watt <JPEWhacker@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Fri, 3 Aug 2018 14:01:12 +0000 (15:01 +0100)]
cmake: fix linking in multilib SDKs
CMake, in it's infinite wisdom, when searching for libraries (using
find_library) will look inside the libraries PREFIX/lib, PREFIX/lib32, and
PREFIX/lib64. In a multilib SDK more than one of these will exist (potentially
all three) and obviously for a given build configuration only one of those is
valid. This search path is hard-coded deep inside CMake but by setting
CMAKE_FIND_LIBRARY_CUSTOM_LIB_SUFFIX we can bypass it and set an explicit
lib<SUFFIX> to use.
Do this by writing ${baselib} into the environment file, and then stripping
"lib" from this to obtain the suffix in the cmake-specific environment script,
which is then read in the CMake toolchain.
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Chen Qi [Thu, 2 Aug 2018 04:52:43 +0000 (12:52 +0800)]
man-db: remove '--disable-cache-owner' option
The following error appeared at boot.
systemd-tmpfiles[115]: [/usr/lib/tmpfiles.d/man-db.conf:1] Unknown user '1w'
By default cache owner is enabled and defaults to 'man'. Users could
supply '--enable-cache-owner=[ARG]' to change the default cache owner.
Using '--disable-cache-owner' leaves the ownership of system-wide
cache files unconstrained, and users will allowed to modify them.
We'd better keep the default behavior, just like other distros do.
I can guess that we used '--disable-cache-owner' to bypass the following
error at do_install.
| chown: invalid user: ‘man:man’
The 'man' user is provided by base-passwd recipe, so add it to DEPENDS.
Signed-off-by: Chen Qi <Qi.Chen@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Maxin B. John [Tue, 31 Jul 2018 14:15:45 +0000 (17:15 +0300)]
machine: remove kernel-modules from RRECOMMENDS
qemumips,qemumips64 and x86 based machines include kernel-modules by
default. Remove the kernel modules from RRECOMMENDS to make it consistent
across the machines.
Tested using core-image-sato on qemumips, qemumips64, qemux86 and
qemux86-64
[YOCTO #12383]
Signed-off-by: Maxin B. John <maxin.john@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
u-boot attempts to build a Python library called pylibfdt. By default,
u-boot would attempt to use the build host's Python interpreter, which
causes numerous problems, not least of which is that it fails if the
host doesn't have the Python development package installed (complaining
about not being able to find Python.h)
Rectify this situation by including the proper build time dependencies
for pylibfdt and passing the proper arguments to make.
[YOCTO #12867]
Signed-off-by: Joshua Watt <JPEWhacker@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Tue, 31 Jul 2018 17:05:41 +0000 (18:05 +0100)]
python[3]native: allow use of both native Python 2 and 3
Instead of both classes using ${PYTHON_PN} (which is set by python-dir and
python3-dir), hardcode the right values so that it is possible to inherit both
pythonnative and python3native at the same time.
If both are inherited then both native Pythons are available, but the last to be
inherited will be the one to set PYTHON.
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Tue, 31 Jul 2018 17:05:40 +0000 (18:05 +0100)]
meson: don't dump full error log on failure
If the configure fails then we don't really want to see hundreds of lines of
test output (this would be similar to dumping out autoconf's config.log). The
error includes the path of the full log if further debugging is required.
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Andre McCurdy [Tue, 31 Jul 2018 13:57:56 +0000 (06:57 -0700)]
openssl: fix missing dependency on hostperl-runtime-native
Openssl 1.1 requires perl in order to build (just as openssl 1.0
does). The missing dependency has gone unnoticed up to now since
hostperl-runtime-native is included in ASSUME_PROVIDED.
Signed-off-by: Andre McCurdy <armccurdy@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
