Khem Raj [Tue, 30 Mar 2021 21:30:19 +0000 (14:30 -0700)]
libgcc-initial: Do not build fp128 to decimal ppc functions
These functions depend on glibc headers e.g.
stdlib.h/fenv.h/string.h/stdio.h which do not exist when building
lbgcc-initial, and these functions are not needed during glibc build
so we are fine to disable them
certain applets are enabled but the long options are not enabled for
them, it results in subtle failures in ptests where its expecting these
options e.g. gzip --best is commonly used in many package tests e.g.
This patch add all the long options to this fragment. The long options
for a tool will only get enabled if the corresponding tool/feature is
enabled in main defconfig, otherwise it will be ignored in final .config
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Khem Raj [Mon, 3 May 2021 06:36:00 +0000 (23:36 -0700)]
cml1.bbclass: Return sorted list of cfg files
Since these files are passed to merge_config.sh its better to use a
sorted list, so we can be sure that this list is always fed in same
order irrespective of python versions on host
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
rootfs.py: find .ko.gz and .ko.xz kernel modules as well
* with xz PACKAGECONFIG enabled in kmod and xz module compression enabled in kernel
the do_rootfs task doesn't run depmod in the image, because it thinks there are no modules:
NOTE: No Kernel Modules found, not running depmod
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com> Signed-off-by: Christophe Chapuis <chris.chapuis@gmail.com> Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When run cyclictest, the parameter enable NUMA. but in some BSP
which not support NUMA, will causes the test program to exit
directly and does not carry out subsequent operations, the
latest changes have fixed this problem. so update to the
latest branch to resolve.
Signed-off-by: Zqiang <qiang.zhang@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
libpam: Provide needed env for tst-pam_start_confdir ptest
tst-pam_start_confdir needs a file called confdir and it should reside
in directory pointed by srcdir env variable, therefore copy confdir into
ptest package and export srcdir before running the ptests
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
103,108d102
< cat: ../y.tab.c: No such file or directory
< cmp: ../y.tab.c: No such file or directory
< cat: /usr/lib/bash/ptest/config.h: No such file or directory
< cmp: /usr/lib/bash/ptest/config.h: No such file or directory
< cat: /usr/lib/bash/ptest/version.h: No such file or directory
< cmp: /usr/lib/bash/ptest/version.h: No such file or directory
FAIL: run-heredoc
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Sat, 1 May 2021 11:25:55 +0000 (12:25 +0100)]
pybootchart/draw: Avoid divide by zero error
When disk stats don't run frequenctly enough, we see divide by zero
errors. The code already has a fallback path so ensure we use it
for this case too.
[YOCTO #14360]
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Khem Raj [Sat, 1 May 2021 21:12:21 +0000 (14:12 -0700)]
pkgconfig: Fix nativesdk builds for mingw sdk hosts
pkgconfig uses a vendored version of ancient glib-2.0, which does not
have many of fixes that current glib-2.0 will have, we enable this
internal version for nativesdk/native recipe, which on mingw hosts does
not work well, as its missing necessary mingw support. This patch
backports couple of fixes which makes GCC11 happy
but its going to be a constant source of pain as long as we support mingw
Paul Barker [Wed, 28 Apr 2021 09:51:32 +0000 (10:51 +0100)]
prservice: Use new connect API
The new prserv connect() function decouples the code in oe-core from the
exact classes and implementation details used within bitbake. This
allows us to more easily switch over to a new asyncrpc based prservice.
Signed-off-by: Paul Barker <pbarker@konsulko.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This will help identifying packages using __TIME__, __DATE__ and
__TIMESTAMP__ macros in code, its a warning from the compiler and the
packages which use -Werror will break the build with this and where they
don't we will atleast have a warning in the build logs
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
gcc-runtime: Fix __FILE__ related reproducablity issues
libstdc++ uses assertion macros which use __FILE__ macros and
if (__builtin_expect(!bool(_Condition), false)) \
std::__replacement_assert(__FILE__, __LINE__, __PRETTY_FUNCTION__, \
#_Condition)
This ends up using absolute paths into build tree for the cases where
the charconv header is used, therefore replace the file prefix paths
with on-target paths to make them build dir independent
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Petr Vorel [Wed, 28 Apr 2021 19:17:36 +0000 (21:17 +0200)]
ltp: Replace musl patches with do_patch[postfuncs]
MUSL related build fixes are not going to be upstreamed. They just not
compile broken files, thus replace them with upstream solution for CI:
just deleting files for musl (easier to maintain).
Signed-off-by: Petr Vorel <petr.vorel@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Sat, 1 May 2021 16:10:21 +0000 (17:10 +0100)]
patchelf: Fix note section alignment issues
Improve note section normalization was added to patchelf in recent versions
however if fails if there are two note sections which aren't sized to match
section alignment. Tweak the code to account for section alignment.
This fixes patchelf failures on the autobuilder, particularly to ccache-native.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Fri, 30 Apr 2021 09:35:02 +0000 (10:35 +0100)]
libxcrypt: Update to 4.4.19 release and fix symbol version issues
This patch upgrades to the 4.4.19 release and replaces a configure patch
from "libxcrypt: fix sporadic failures in nativesdk-libxcrypt-compat" with
a fix to avoid leading spaces in CFLAGS causing failures.
The license changed a few filenames listed in the license but the overall
license remains unchanged.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
buildstats.bbclass: collect data in the same file.
Previously "at interval" and "on failure" logs were collected
in separate files. Collect both types of logging in the same
file for better analysis.
Introduced new variable which allows different set of commands
to be run by the different logging, interval or failure. The
variables are BB_LOG_HOST_STAT_CMDS_INTERVAL and
BB_LOG_HOST_STAT_CMDS_FAILURE respecteviely.
Signed-off-by: Sakib Sajal <sakib.sajal@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This is the first step towards rebasing the AB matrix from sato to
weston; the eventual goal is to keep sato only in core-image-sato
image.
The broader rationale is that X11 is effectively deprecated technology
at this point with only minimal maintenance; standalone X server will not
be developed any further, and all attention currently is towards making
it work well under Wayland.
I believe YP should be defaulting to Wayland and not X11.
Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This adds support for the Qemu Machine Protocol [0] extending
the current dump process for Host and Target. The commands are
added in the testimage.bbclass.
Currently, we setup qemu to stall until qmp gets connected and
sends the initialization and continue commands, this works
correctly. If the UNIX Socket does not exist, we wait an timeout
to ensure to socket file is created.
With this version, the monitor_dumper is created in OEQemuTarget
but then set in OESSHTarget as that's where we get the SSH failure
happens. Python's @property is used to create a setter/getter type
of setup in OESSHTarget to get overridden by OEQemuTarget.
By default the data is currently dumped to files for each command in
TMPDIR/log/runtime-hostdump/<date>_qmp/unknown_<seq>_qemu_monitor as
this is the naming convenstion in the dump.py code.
We use the qmp.py from qemu, which needs to get installed in the
recipe-sysroot-native of the target image.
The qmp python module supports the Qemu Machine Protocol [0].
This module needs to be installed in a known location so the
qemurunner python script can find the qmp module.
This change causes it to be installed in the recipe-sysroot-native
of the target image and that directory can be added to the python
sys.path that needs to use the qmp.py module.
gcc-cross-canadian: Install LTO linker plugin to BFD searchable location
This helps binutils provided tools ar/ranlib/nm to find the LTO linker
plugin automatically as well which makes it equivalent to gcc-ar/gcc-nm/gcc-ranlib
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
gcc-cross: Install linker LTO plugin for binutils tools
This will ensure that ar/ranlib/nm can load the lto linker plugin like
gcc-ar, gcc-nm, gcc-ranlib does, this will let the behaviour match
between gcc wrappers for these tools, this should help LTO builds for
packages
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
gcc-target: Create a LTO plugin symlink in bfd-plugins directory
This directory is scanned by binutils provided ar,ranlib,nm for plugins
that it can load automatically, putting liblto_plugin.so in their means
we do not need gcc-ar, gcc-nm, gcc-ranlib particularly as normal
ar/ranlib/nm tools will work equally well as they can now use this
linker plugin by default
This also mean we can revert back to using ar/ranlib/nm as default
providers for AR/NM/RANLIB on target
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Marek Vasut [Fri, 12 Mar 2021 20:57:36 +0000 (21:57 +0100)]
linux-firmware: Package RSI 911x WiFi firmware
The RSI 911x WiFi firmware is already part of the linux-firmware
repository, package it to make it easily available.
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Richard Purdie <richard.purdie@linuxfoundation.org> Cc: Steve Sakoman <steve@sakoman.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Sat, 24 Apr 2021 14:23:54 +0000 (15:23 +0100)]
yocto-check-layer: Avoid bug when iterating and autoadding dependencies
If iterating a layer with multiple components and auto-adding dependencies
the tests can break since layers are never removed and order isn't guaranteed
to account for that.
Fix this by resetting the layer list back to the original list each time
before auto-adding the dependencies in each case.
This fixes scanning of meta-openembedded in particular where the sublayers
may not be added in order of minimal dependency.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Mikko Rapeli [Thu, 22 Apr 2021 14:41:14 +0000 (17:41 +0300)]
unzip: use optimization from bitbake
Build with bitbake default optimizations, e.g. O2,
instead of custom O3. Reduces unzip binary package
size from 304401 to 283921 bytes, and enables building
with Os to even further reduce binary size if needed
and configured for the whole system.
Signed-off-by: Mikko Rapeli <mikko.rapeli@bmw.de> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Kai Kang [Fri, 23 Apr 2021 03:29:04 +0000 (11:29 +0800)]
kernel-yocto.bbclass: chdir to ${WORKDIR} for do_kernel_checkout
It chdirs to ${S} at the beginning of task do_kernel_checkout. Then it
removes ${S} when it still resides in ${S}. It may fail to run the task
do_kernel_checkout when bitbake is called by third-part wrapper script.
So chdir to ${WORKDIR} by default for do_kernel_checkout. And it will
chdir to ${S} afterwards in task do_kernel_checkout.
Signed-off-by: Kai Kang <kai.kang@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Mikko Rapeli [Thu, 22 Apr 2021 14:41:52 +0000 (17:41 +0300)]
lz4: use CFLAGS from bitbake
Currently lz4 uses it's own defaults which include O3 optimization.
Switch from O3 to bitbake default O2 reduces binary package size
from 467056 to 331888 bytes. Enables also building with Os if needed.
Signed-off-by: Mikko Rapeli <mikko.rapeli@bmw.de> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* aad50fcd fix regression in dl_iterate_phdr reporting of modules with no TLS
* 0ea78a64 nscd: fall back gracefully on kernels without AF_UNIX support
* 95a540e1 mallocng/aligned_alloc: check for malloc failure
* 2c00f95c make epoll_[p]wait a cancellation point
* 521b4d27 fix dl_iterate_phdr dlpi_tls_data reporting to match spec
* 122002f0 remove no-longer-needed special case handling in popen
* 8ef9d46f use internal malloc for posix_spawn file actions objects
* cfdfd5ea don't fail to map library/executable with zero-length segment maps
* e48e99c1 suppress isascii() macro for C++
* b129cd86 guard against compilers failing to handle setjmp specially by default
* 3309e2d7 aarch64/bits/mman.h: add PROT_MTE from linux v5.10
* 44331150 aarch64/bits/hwcap.h: add HWCAP2_MTE from linux v5.10
* 42aa19a0 add aarch64/bits/mman.h with PROT_BTI from linux v5.8
* b7554b5e aarch64/bits/hwcap.h: add HWCAP2_BTI from linux v5.8
* 87b8f148 signal.h: add MTE specific SIGSEGV codes from linux v5.10
* 19239cde sys/prctl.h: add MTE related constants from linux v5.10
* 8b29f023 elf.h: add NT_ARM_TAGGED_ADDR_CTRL from linux v5.10
* d7210f0c sys/mman.h: add MAP_HUGE_16KB from linux v5.10
* a7456524 sys/mount.h: add MS_NOSYMFOLLOW from linux v5.10
* 54ca1cc7 sys/membarrier.h: add new constants from linux v5.10
* fd285f9d bits/syscall.h: add process_madvise from linux v5.10
* 49b6df3d fix error return value for cuserid
* cc577d0e fix misuse of getpwuid_r in cuserid
* a75283d7 cuserid: don't return truncated results
* ef137da6 cuserid: support invocation with a null pointer argument
Was not able to reproduce the reported non-deterministic failure.
(ran 20 times on qemux86-64 on a relatively isolated machine)
it might be related to the AB-INT issues,
but it seems to only affect ARM builders now.
Also no action taken by upstream valgrind to fix
https://bugs.kde.org/show_bug.cgi?id=430321
Let it run on AB to see if failure was fixed by uprev to 3.17.0.
if not, we can gather more data from the AB failure.
[YOCTO #14051]
Signed-off-by: Yi Fan Yu <yifan.yu@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
