Add PACKAGECONFIG lksctp to fix qa warning:
| WARNING: QA Issue: iperf3 rdepends on lksctp-tools, but it isn't a
| build dependency? [build-deps]
Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
1. inherit systemd and move the systemd unit files to correct dir
2. inherit allarch
3. add bash/python to RDEDEPENDS
Signed-off-by: Roy Li <rongqing.li@windriver.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
When compile with gcc 5.x, libhugetlbfs fails with:
| hugeutils.c: In function '__lh_hugetlbfs_setup_env':
| hugeutils.c:304:40: error: expected identifier or '(' before restrict'
| char *p, *tok, *exe, buf[MAX_EXE+1], restrict[MAX_EXE];
| ^
Rename variable to avoid using keyword restrict as variable name to fix it.
Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
There was a patch fixing the same issue:
libhugetlbfs-avoid-search-host-library-path-for-cros.patch
but missed the elf_i386.xB*, this patch fixes for it.
Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Make fio's NUMA support optional via PACKAGECONFIG.
Enable this feature by default, except on ARM where libnuma/numactl
is not available.
Signed-off-by: Dominic Sacré <dominic.sacre@gmx.de>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
The tool has support for multiple backends: X11, drm and Wayland.
This demanded to a fix, which has been sent upstream, to make the
flavor configuration predictable and the recipe has been rework to
support them all, making the PACKAGECONFIG to respect the
DISTRO_FEATURES available by default.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
The development has been moved for github thus it makes sense to use
this, instead of the tarball.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
WAWARNING: QA Issue: fio rdepends on numactl, but it isn't a build dependency? [build-deps]
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Fixes:
| WARNING: QA Issue: fio rdepends on zlib, but it isn't a build dependency? [build-deps]
Signed-off-by: Dominic Sacré <dominic.sacre@gmx.de>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
COMPATIBLE_HOST included support for x86_64 but not for
32-bit x86 architectures - added i586 arch name to cover qemux86, etc.
Signed-off-by: Gary S. Robertson <gary.robertson@linaro.org>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
If you want to use iperf3 to measure your board's IP performance, you
may also want to compile if for the build host, because unlike iperf,
iperf3 isn't readily available as a standard Ubuntu package for example.
Add a BBCLASSEXTEND="native" to the recipe so that you just can build
iperf3-native and have bitbake compile it for you, instead of having to
download, compile and install it manually on the build machine.
Signed-off-by: Mike Looijmans <mike.looijmans@topic.nl>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
iperf3 is the new successor of iperf. The project is now mature enough
as it started in 2009 and has stable releases. It supports reporting
results in the portable JSON format, provides more information than
iperf does and also adds some features from other tools such as nuttcp
and netperf that were missing from the original iperf.
This recipe is based on the iperf v3.0.8 recipe posted to the list by
Guy Morand
(http://lists.openembedded.org/pipermail/openembedded-devel/2014-October/098403.html)
with the following changes:
- Pass the 'foreign' option to automake.
- The Makefiles for the examples only work properly when the build is
done in the same directory as the source. Fix this in Makefile.am and
run bootstrap.sh to regenerate Makefile.in.
- Bump the version to the latest stable release.
Signed-off-by: Ben Shelton <ben.shelton@ni.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
The license is not a common only and seems to be a 'free distribution'
license but without explicit rules about modification or similar. Due
that, we copied the license text from the README file and added it
onto nbench-byte license file.
This fixes following QA error:
,----
| WARNING: Recipe nbench-byte, LICENSE (freely distributable) has
| invalid format, LICENSES must have operator "([&|() ])" between
| them.
`----
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
The configure step uses a package header to check for support
of types/declares, this package header encapsulates all the
required system headers for providing the needed resources.
When configured from an out of tree directory the package
header is not found due to the hard path.
We now make the path for package header relative to 'srcdir' so
it is found appropriately.
Signed-off-by: Awais Belal <awais_belal@mentor.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Fixed computation of page size and text segment offset for various arm
architectures - including both LE and BE variants of armv7 as well as
aarch64
Upstream Status: Accepted at libhugetlbfs project
Signed-off-by: Gary S. Robertson <gary.robertson@linaro.org>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
In a non-native cross-development scenario, the Makefile
only recognized i386 or x86_64 PC architectures.
Extended this to also recognize i486, i586, and i686
Upstream Status: Accepted but not yet applied by libhugetlbfs project
Signed-off-by: Gary S. Robertson <gary.robertson@linaro.org>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
This solves the following warning:
libhugetlbfs-2.18: libhugetlbfs-tests requires /bin/bash, but no providers in
its RDEPENDS [file-rdeps]
libhugetlbfs-2.18: libhugetlbfs requires /bin/bash, /usr/bin/perl, but no
providers in its RDEPENDS [file-rdeps]
Signed-off-by: Chong Lu <Chong.Lu@windriver.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
This is the package recipe for the phoronix test suite v5.2.1. The
tool is for benchmarking and testing of the system.
Signed-off-by: Todd Brandt <todd.e.brandt@linux.intel.com>
----
meta-oe/recipes-benchmark/phoronix-test-suite/phoronix-test-suite_5.2.1.bb | 24 ++++++++++++++++++++++++
1 file changed, 24 insertions(+)
v2: changes based on Ross Burton's comments. However, I kept this line:
S = "${WORKDIR}/phoronix-test-suite", since the tarball folder doesn't
include the version number.
v3: changed the license file locatoin and moved the RDEPENDS field down
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Perl:
bonnie-scripts/usr/bin/bon_csv2html:#!/usr/bin/perl
bonnie-scripts/usr/bin/bon_csv2txt:#!/usr/bin/perl
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
- I don't see any evdidence that lmbench needs debianutils to run:
debianutils provides:
run-parts, tempfile, savelog, which, ischroot,
add-shell, installkernel, remove-shell
but none of them is used by lmbench scripts and I tested it work
fine without debianutils, so remove debianutils from RDEPENDS
- Some scripts in lmbench are perl scripts, so add
RDEPENDS on perl (no addtional perl module is needed).
Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
lmbench can't proceed on some targets. The memory check fails because the
memory latency of each page is longer then 10us, which is a time limit set
in the original memsize.c.
The memory latency is very different on different targets due to the
hardware and current system load. The targets with slower memory
chips or heavy system load need much longer time to read or write
the memory.
This fix changes the fixed time limit of 10us to a specific value calculated
from the runtime target.
Also set an upper limit of memory size used for lmbench testing. The memory
check sometimes fails if the target has a large amount of memory, for
example more than 4G.
Signed-off-by: Qingming Su <qingming.su@windriver.com>
Signed-off-by: Fupan Li <fupan.li@windriver.com>
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Upstream Status: Submitted
This patch applies the same technique used by Koen Kool in the following patch
which was accepted by the libhugetlbfs project:
[0a4f6] Add aarch64_be_support 2014-03-31 10:52:37
It modifies the libhugetlbfs Makefile to mark all 32-bit arm architectures
as supported by the libhugetlbfs build. Builds and successful functional
tests have been performed for armv7a LE and BE runtime platforms.
This patch replaces and renders obsolete the following patch:
arm32-support.patch by: Chunrong Guo <B40290@freescale.com>
Signed-off-by: Gary S. Robertson <gary.robertson@linaro.org>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
qemux86_64 has the same issue with powerpc64, set
BUILDTYPE=NATIVEONLY as default
Fix the below build issue for 64bit platforms:
| fatal error: gnu/stubs-32.h: No such file or directory
| # include <gnu/stubs-32.h>
| ^
| compilation terminated.
| make: *** [obj32/elflink.o] Error 1
Signed-off-by: Ting Liu <b28495@freescale.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
If obj32 needed, user can build lib32-libhugetlbfs.
Fix the below build issue for 64bit powerpc boards:
| fatal error: gnu/stubs-32.h: No such file or directory
| # include <gnu/stubs-32.h>
| ^
| compilation terminated.
| make: *** [obj32/elflink.o] Error 1
Signed-off-by: Ting Liu <b28495@freescale.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
In recent libhugetlbfs source, powerpc/powerpc64 is changed back to
ppc/ppc64 (to match with the output of `uname -a`).
Update the recipe to avoid build issue for now.
Signed-off-by: Ting Liu <b28495@freescale.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Libhugetlbs doesn't generate proper .so.X libs, only two .so files. Package those in ${PN} and silence the resulting QA warning. This fixes:
ERROR: QA Issue: libhugetlbfs-tests rdepends on libhugetlbfs-dev
Signed-off-by: Koen Kooi <koen.kooi@linaro.org>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Previous location is no longer accessible and breaks builds that include
cpuburn-neon. This updates the location to fix the build breaks
Signed-off-by: Lauren Post <lauren.post@freescale.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Fixed SRC_URI:
* ${PN} -> ${BPN}, use ${BP} if it was ${PN}-${PV}
* ${P} -> ${BP}
Otherwise we would meet do_fetch errors when we do the multilib, native
or nativesdk build.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
This extends the dependencies on pointer.h to fix following error:
| ./sysinfo.sh arm-poky-linux-gnueabi-gcc ...
| In file included from misc.h:36:0,
| from misc.c:23:
| nmglobal.h:29:21: fatal error: pointer.h: No such file or directory
| #include "pointer.h"
| ^
| compilation terminated.
| make: *** [misc.o] Error 1
| make: *** Waiting for unfinished jobs....
| ERROR: oe_runmake failed
| WARNING: exit code 1 from a shell command.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
This fixes the Makefile so the client.txt file is copied from the
original source tree. This fixes the out of tree build error.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
* Update recipe:
- update SRCREV to latest release, 2.18
- switch back to master branch
- increase verbosity (pass V=1 to EXTRA_OEMAKE) to make debugging easier
* Drop patches:
- fix-lib64-can-not-be-shiped-in-64bit-target.patch
replaced by LIBARGS definition. the build system allows to override libdir
using LIB32 and LIB64 variables
- 0001-install-perl-lib-to-directory-perl-instead-of-perl5.patch
replaced by a one-liner to fixup perl module directory hardcoded to perl5
* Refresh skip-checking-LIB32-and-LIB64-if-they-point-to-the-s.patch
- only disable the error to skip the check
* Add patches:
- 0001-aarch64-fix-cross-compilation.patch - submitted upstream
allow to override CC and use it for aarch64 case like the other arch
- 0001-aarch64-fix-page-size-not-properly-computed.patch - submitted upstream
fix typo on MB variable (MB -> $MB) and add the missing linker variants
(aarch64elfb, aarch64linux, aarch64linuxb)
- 0001-replace-lib-lib64-hardcoded-values-by-LIBDIR32-LIBDI.patch
submitted upstream. replace lib/lib64 hardcoded values by LIBDIR32 and
LIBDIR64 variables
* Remove patches that aren't applied in the recipe (they're all merged
upstream and available in 2.18 release):
- libhugetlbfs/files/aarch64-support.patch
- aarch64-unit-test-fixes.patch
- add-PROT-NONE-to-the-mprotest-test.patch
Signed-off-by: Fathi Boudra <fathi.boudra@linaro.org>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Kai Kang
Roy Li
Chunrong Guo
Otavio Salvador
Soren Brinkmann
Jianchuan Wang
Dominic Sacré
Martin Jansa
Josep Puigdemont
Yi Zhao
Armin Kuster
gary.robertson@linaro.org
Andre McCurdy
Raphael Silva
mike.looijmans@topic.nl
Ben Shelton
Belal, Awais
Chong.Lu@windriver.com
Todd E Brandt
Robert Yang
Yasir Khan
Jackie Huang
Ting Liu
Koen Kooi
Lauren Post
Diego Rondini
Fathi Boudra