This array will be re-used in another function in a follow-up commit,
so it makes sense to factor it out.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
The remove_extra_deps() function removes dependencies that we are not
interested in seeing in the dependency graph. It does this for all
packages, except the 'all' package, which on full dependency graphs is
the root of the tree.
However, this doesn't take into account package-specific dependency
graphs (i.e make <pkg>-graph-depends) where the root is not 'all', but
'<pkg>'. Due to this, dependencies on "mandatory deps" were not
visible at all, i.e the toolchain package (and its dependencies) and
the skeleton package (and its dependencies) were not displayed in
package-specific dependency graphs.
To fix this, we use the existing rootpkg variable instead of
hardcoding 'all'.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Currently, we avoid drawing the dependencies that we call 'target
exceptions', becasue they initially were returned by 'show-targets',
when they in fact were not really packages and thus should not be on
the graph.
However, those two exceptions have no longer been reported in the output
of show-targets since we merged very old initial top-level parallel
build way back in 2014, with commit a24877586a (Makefile: add support
for top-level parallel make), where they had been converted into purely
internal rules.
4 years have passed, we can now drop those exceptions from the
graph-depends script.
This concludes the cleanup initiated three years ago with commit
0b32791f00 (graph-depends: remove absent targets from
TARGET_EXCEPTIONS).
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
His e-mail @rockwellcollins.com is bouncing:
<bryan.brinsko@rockwellcollins.com>: host
mail05.rockwellcollins.com[205.175.227.28] said: 550 #5.1.0 Address
rejected. (in reply to RCPT TO command)
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
As reported in bug #11426, the ppsfind shell script uses /bin/bash,
but the Buildroot pps-tools package doesn't depend on bash. In fact,
upstream has fixed the problem, and the script can now be used with a
POSIX shell, and the shebang is /bin/sh.
This commit therefore bumps pps-tools to the latest upstream commit,
which is precisely this fix.
Fixes bug #11426.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
The Lynx configure script uses pkg-config when available:
checking for nios2-buildroot-linux-gnu-pkg-config... /home/thomas/projets/buildroot/output/host/bin/pkg-config
checking pkg-config for openssl... yes
[...]
checking pkg-config for ncurses... yes
Using pkg-config avoids build failures such as:
checking for _nc_freeall... no
configure: error: Configuration does not support color-styles
make: *** [/home/test/autobuild/run/instance-1/output/build/lynx-2.8.9rel.1/.stamp_configured] Error 1
When building with "make lynx", so that pkg-config is not built
before. The issue is that in this case, lynx configure script picks up
the ncurses6-config script for the host ncurses instead of the one in
staging. Using pkg-config solves that nicely.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
When using an initramfs, on the first-pass build, we create a dummy cpio
so that the build succeeeds. The real cpio will come later, and we'll do
a second-pass build to use the actual cpio.
However, when we touch that dummy cpio, the images/ directory may not
yet exist, since commit d0f4f95e39 (Makefile: rework main directory
creation logic) removed its creation at the begining of the build, to
only at the moment we need it, i.e. during the *_INSTALL_IMAGES_CMDS
steps.
However, the linux build is not a _INSTALL_IMAGES_CMDS step, so there is
no guarantee that images/ already exist at that time.
Fix that by explicitly creating images/ before touching the dummy cpio.
Reported-by: Matthew Weber <matthew.weber@rockwellcollins.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Matthew Weber <matthew.weber@rockwellcollins.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Reviewed-by: Matthew Weber <matthew.weber@rockwellcollins.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
If BR2_TOOLCHAIN_HAS_SYNC_4 is not set, define -DNO_ATOMICS=1 to
fallback on lock
Fixes:
- http://autobuild.buildroot.org/results/7b3bc71d64aedf617dde35798099e16d7ae46713
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
[Thomas: invert condition.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
semaphore.h is available only if HAVE_LIB_PTHREAD and
HAVE_SEM_POSIX are defined
Fixes:
- http://autobuild.buildroot.org/results/1c95898b2833683a22bbe2ff8471fa08d94210e1
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
[Thomas: update upstream status of the patch.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
libiscsi uses cunit for its test suite, and autodetects its presence.
Usually, we just try to disable tests altogether, but there is no way to
do so. So, ensure proper ordering.
Note: there is an ac_cv_have_cunit variable, but it is not a true
ac_cv* cache variable, and the value provided on the command line or
environment is properly ignored.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
The pre-rendered, bundled ones are still installed, though, but they
get removed in target-finalize anyway.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Stack protector has been added in version 3.0.3 by
2268c27754
However, some compilers could missed the needed library (-lssp or
-lssp_nonshared) at linking step so use ax_check_link_flag instead of
ax_check_compile_flag
Fixes:
- http://autobuild.buildroot.org/results/0b90e7dca2984652842832a41abad93ac49a9b86
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
KEY_ALS_TOGGLE has been added in kernel 4.8 with
9ee2748712
So add it to missing.h to fix build with kernel older than 4.8
Fixes:
- http://autobuild.buildroot.org/results/0c0aff02df91cdb869efa01e397f7ccc0d9f69af
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Reviewed-by: Matthew Weber <matthew.weber@rockwellcollins.con>
[Thomas: update upstream status.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Allow ejabberd user to run ejabberdctl.
Signed-off-by: Johan Oudinet <johan.oudinet@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Commit c2ea056a1b (package/qt5/qt5base: use ccache for building host
code) introduced a tweak to qt5's mkspec file, to define the compilers
(C and C++) to use to build qmake.
In doing so, it changed the variables in that file, from this layout:
VAR = value
to this:
VAR=value
During its configuration phase, qt5 will parse that file using the
system awk.
However, the system awk is not necessarily GNU awk; some systems,
especially Debian ones, use mawk as the default awk implementation.
mawk does not behave the same as GNU awk, and this causes qt5 to extract
empty values, and thus leads to build failures.
Fix that by replacing the variables with similar layout, keeping the
spaces around the equal signs.
Fixes:
http://autobuild.buildroot.org/results/cb5/cb555a124bb3bdb9c5a3465673c21022d94cf2ca/
Reported-by: Alexey Brodkin <Alexey.Brodkin@synopsys.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Alexey Brodkin <Alexey.Brodkin@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Acked-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
In commit cd9d58f1fc ("toolchain: bumb
ARC tools to arc-2018.09 release"), the ARC-specific binutils, gcc and
gdb versions were updated to use the 2018.09 release. However, they
are mistakenly pointing to a branch rather than a tag. This commit
fixes that by using the proper release tag.
Signed-off-by: Evgeniy Didin <Evgeniy.Didin@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: arc-buildroot@synopsys.com
[Thomas: rework commit log]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
The static devices defined by packages are currently added to the full
device table when two conditions are met:
(1) ROOTFS_DEVICE_TABLES is non-empty
(2) BR2_ROOTFS_DEVICE_CREATION_STATIC=y
(2) is obviously correct. However, depending on (1) is not correct: if
the user doesn't provide any custom permission table and custom device
table, then ROOTFS_DEVICE_TABLES will be empty.
So instead, move the addition of the package-defined static devices
outside of condition (1), and have it only under condition (2).
Reported-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
In commit 6b50f988ad ("fs/common.mk:
rename internal variable"), USERS_TABLE was renamed to
ROOTFS_FULL_USERS_TABLE.
This commit follows the same direction by renaming the
FULL_DEVICE_TABLE variable to ROOTFS_FULL_DEVICE_TABLE.
In addition, for consistency, the file itself is renamed
full_device_table.txt.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
This tool allows to verify if the OpenCL environment is set up correctly
and provides information related to the supported OpenCL platforms.
Signed-off-by: Valentin Korenblit <valentin.korenblit@smile.fr>
Signed-off-by: Valentin Korenblit <valentinkorenblit@gmail.com>
[Thomas:
- use the libopencl virtual package
- add LICENSE to the license files]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
We need to install the mesa3d-headers, because the CL headers are not
provided by nvidia-driver (just like it didn't provide the OpenGL
headers).
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Within the BR2_PACKAGE_NVIDIA_DRIVER_XORG condition, some "="
assignements are used for various variables, which are also appended
in other conditions below in nvidia-driver.mk.
It works fine because those assignements appear earlier in the .mk,
but it is a bit error-prone, so let's use += when adding values to
those variables.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
This patch provides Clover, the OpenCL 1.1 API implementation by Mesa
for AMD GPUs. It generates libOpenCL.so.
Add --disable-opencl-icd because in Mesa 18 it defaults to on. When
disabled, the shared library is named libOpenCL instead of
libMesaOpenCL and CL headers are installed.
Given that clc headers are being installed to a non-standard location,
it is necessary to specify this path in configure.ac. Otherwise,
pkg-config will output the absolute path to these headers located in
STAGING_DIR, which will cause a runtime error when calling
clBuildProgram.
Signed-off-by: Valentin Korenblit <valentin.korenblit@smile.fr>
Signed-off-by: Valentin Korenblit <valentinkorenblit@gmail.com>
[Thomas:
- improve the description of the patch, based on input from Romain
- register as a libopencl provider]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Valentin Korenblit <a class="moz-txt-link-rfc2396E" href="mailto:valentin.korenblit@smile.fr"><valentin.korenblit@smile.fr></a>
Signed-off-by: Valentin Korenblit <a class="moz-txt-link-rfc2396E" href="mailto:valentinkorenblit@gmail.com"><valentinkorenblit@gmail.com></a>
Signed-off-by: Thomas Petazzoni <a class="moz-txt-link-rfc2396E" href="mailto:thomas.petazzoni@bootlin.com"><thomas.petazzoni@bootlin.com></a>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
A flexible forms validation and rendering library for
Python web development.
https://wtforms.readthedocs.io/
Signed-off-by: Grzegorz Blach <grzegorz@blach.pl>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
This patch allows to select the installation of additional commands
which are part of the sunxi-tools. It's now possible to e.g. install
sunxi-fel on the target device. The corresponding options have been
added to Config.in and sunxi-tools.mk has been modified respectively.
The default setting is to only build sunxi-nand-part.
On the host building of the misc-tools target is added, which provides
sunxi-nand-image-builder and phoenix_info.
Signed-off-by: Alex Kaplan <kaplan2539@gmail.com>
[Thomas:
- properly format Config.in
- do not select BR2_PACKAGE_HOST_LIBUSB in Config.in.host, since this
option doesn't exist
- properly indent code in sunxi-tools.mk]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Commit e13855c48f wrongly added
depends on BR2_TOOLCHAIN_USES_UCLIBC && !BR2_USE_MMU
to display the comment "libbsd needs a toolchain w/ threads, wchar"
The same error has also been made for minizip.
To fix this issue, move dependency
!(BR2_TOOLCHAIN_USES_UCLIBC && !BR2_USE_MMU) under
BR2_PACKAGE_LIBBSD_ARCH_SUPPORTS.
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Both p1_oauth2 and jiffy are runtime dependencies. Mark the
corresponding select in the Config.in file with a # runtime
comment.
Signed-off-by: Johan Oudinet <johan.oudinet@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Currently all random defconfigs which are used in autobuilder use size
optimizaion (-Os), since BR2_OPTIMIZE_S=y is the default.
Adding "-O2" optimization will give better test coverage.
In many cases software gets built with speed optimization rather than
size optimization. So let's add Level 2 optimizaion option to be
generated in random defconfigs, so we could be able to test how
packages are built with "-O2" in autobuilder.
Signed-off-by: Evgeniy Didin <Evgeniy.Didin@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: arc-buildroot@synopsys.com
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
In preparation of more renames, rename the variable that points to the
final users table.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
[Thomas: as suggested by Arnout, use ROOTFS_FULL_USERS_TABLE instead
of ROOTFS_FINAL_USERS_TABLE.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
libcurl doesn't find any trust path for CA certs when it cross-compiles.
When using OpenSSL, it is explicitly configured to use the SSL cert
directory with OpenSSL style hash files in it. But with GnuTLS, it gets
nothing.
Rather than configure libcurl to use the OpenSSL directory or a bundle
file, configure it to use the GnuTLS default. This way the CA certs
path can be configured in one place (gnutls) and then libcurl and anyone
else who uses gnutls can default to that.
Also, when libcurl with gnutls is configured to use a directory, it ends
up loading each cert three times.
Signed-off-by: Trent Piepho <tpiepho@impinj.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Gnutls is building with no default location to look for CA certs. Since
there are buildroot packages to provide these, configure it to use them
by default.
Configure gnutls to find them using the bundle file which contains all
certs, rather than looking in the cert directory. When gnutls is told
to use the directory, it loads *every* file in it. This means it loads
the bundle with all certs, then loads each cert a second time using the
individual pem files, and then loads them all the third time via the
hash symlinks to the pem files.
When p11-kit is enabled, use its trust module instead of the bundle
file. p11-kit can be configured to use the bundle (the default), but it
can do other things too, such as integrate with the "trust" command for
adding and removing trust anchors.
Signed-off-by: Trent Piepho <tpiepho@impinj.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Fixes the following check-package warning:
package/docker-cli/docker-cli.mk:1: should be 80 hashes (http://nightly.buildroot.org/#writing-rules-mk)
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Docker upstream has split the Docker daemon and CLI into separate
codebases:
- github.com/docker/engine: daemon, "dockerd" binary
- github.com/docker/cli: "docker" command line interface
This commit splits the docker-engine package into docker-engine and
docker-cli. Conveniently, the Docker project has begun maintaining
two separate release-tagged repositories for the CLI and daemon as of
v18.06-ce-rc1. Previous versions were tagged in a common "docker-ce"
repository which makes compilation awkward for Buildroot, especially
due to some limitations in the new Go package infrastructure.
Docker repositories "engine" and "cli" recently started tagging
releases. Select the latest stable release, v18.09.0.
The CLI is no longer automatically included with the engine. Users
will need to select BR2_PACKAGE_DOCKER_CLI to produce a both docker
and dockerd target binaries.
Docker CLI can be statically compiled. This enables usage of the
system docker client binary to access the parent daemon API from
within containers, where shared libraries are not available.
While at it, drop the useless host-go dependency from docker-engine,
since it's already added by the golang-package infrastructure.
Signed-off-by: Christian Stewart <christian@paral.in>
[Thomas: drop the host-go dependency from both docker-cli and
docker-engine]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
