This commit introduces a virtual package called 'toolchain', located
in 'toolchain/toolchain', which simply depends on
'toolchain-buildroot' or 'toolchain-external' depending on the
selected toolchain backend.
For now, toolchain-buildroot and toolchain-external are still manual
make targets, but the following patches convert those backends to use
the package infrastructure as well.
In addition to this:
* The main Makefile is modified to always make BASE_TARGETS point to
this new toolchain virtual package.
* The main Makefile is changed to include all the toolchain/*/*.mk
files: the toolchain virtual package, and the toolchain-buildroot
and toolchain-external directories.
* The dependency of the toolchain on prepare dirs and dependencies is
moved to the toolchain virtual package. It is moved as a
prerequisite of the "toolchain-source" rule to ensure that all
directories are prepared before we even start extracting the
toolchain.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Luca Ceresoli <luca@lucaceresoli.net>
Acked-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
In order to avoid the work of converting the toolchain-crosstool-ng
logic to the package infrastructure, we remove it from Buildroot,
since it has been deprecated since quite some time.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The 'make clean' recipe is using variables that are not defined without
.config file, causing only a partial cleanup when the .config file is
accidentally deleted.
This patch moves those variables that do not depend on values from .config
outside the BR2_HAVE_DOT_CONFIG check, so that 'make clean' is much more
similar with and without .config.
Since HOST_DIR is determined from BR2_HOST_DIR in .config, the host
directory cannot be cleaned correctly without making assumptions, if no
.config is present. However, to cover most people's use cases, we assume
the default value of $(BASE_DIR)/host in this specific case.
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Since STAGING_DIR is defined as (package/Makefile.in):
STAGING_SUBDIR = usr/$(GNU_TARGET_NAME)/sysroot
STAGING_DIR = $(HOST_DIR)/$(STAGING_SUBDIR)
removing HOST_DIR will automatically remove STAGING_DIR. This patch updates
'make clean' based on this knowledge.
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The 'make help' lines for generating the manual did not consider ePub
correctly. Instead of providing a limitative list for 'make manual', the
text is changed into 'all formats'.
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
The -s option is only mandatory for the charmap,
else the filename is always empty and nothing
is generated.
Signed-off-by: Thierry Bultel <thierry.bultel@wanadoo.fr>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
The $(BUILD_DIR)/.root rule is executed as part of the 'dirs'
target. The 'dirs' target is re-executed at every execution of 'make
external-deps', and make external-deps explicitly tells make to ignore
targets that have already been made (through the -B option). This
means that the $(BUILD_DIR)/.root rule has to be idempotant, which was
not the case this the introduction of the lib32/lib64 symbolic link.
Running 'make external-deps' three times in a row was sufficient to
trigger an error due to symbolic links being incorrectly created. This
patch fixes that.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Similar to how we do it for _PATH / _SYSROOT_DIR. Buildroot gets fairly
confused if the user has set up a custom PKG_CONFIG_LIBDIR, so ensure
it gets unset.
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Make 3.82 no longer sort the result of wildcards (see
http://comments.gmane.org/gmane.comp.gnu.make.bugs/4260). This may break
build reproducibility.
This patch sort results of wildcards to ensure reproducibility.
Signed-off-by: Jérôme Pouiller <jezz@sysmic.org>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
In the past there may have been a very good reason to refer to 'buildroot2', but
these days it's just odd.
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
libpthread.so has a special stripping condition to preserve parts of
it that are needed for debugging. However, due to the usage of 'xargs'
instead of 'xargs -r', the strip command is executed regardless of
whether a libpthread.so file is found or not. This leads to a big
error message being displayed in static-only builds, because strip is
executed without a file argument. Thanks to the '|| true', the build
continues, but still shows a big error message in the middle, which is
not nice.
By using 'xargs -r', we avoid the strip command from being executed
when 'find' doesn't find any match. We also remove the '|| true' to
catch other real errors.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
* Always link lib64 or lib32 to lib
* Only copy the architecture's lib directory to staging
* Also cleanup a couple of mkdirs (concerning some 'lib' directories).
Before this patch:
$ ls -ld host/usr/x86_64-buildroot-linux-gnu/sysroot/{,usr/}lib* target/{usr/,}lib*
drwxr-xr-x 2 samuel users 4.0K Aug 12 22:26 host/usr/x86_64-buildroot-linux-gnu/sysroot/lib/
drwxr-xr-x 2 samuel users 4.0K Aug 12 22:27 host/usr/x86_64-buildroot-linux-gnu/sysroot/lib64/
drwxr-xr-x 5 samuel users 4.0K Oct 30 2012 host/usr/x86_64-buildroot-linux-gnu/sysroot/usr/lib/
drwxr-xr-x 5 samuel users 4.0K Aug 12 22:27 host/usr/x86_64-buildroot-linux-gnu/sysroot/usr/lib64/
drwxr-xr-x 3 samuel users 4.0K Oct 30 2012 host/usr/x86_64-buildroot-linux-gnu/sysroot/usr/libexec/
drwxr-xr-x 2 samuel users 4.0K Aug 12 22:27 target/lib/
lrwxrwxrwx 1 samuel users 3 Aug 12 22:27 target/lib64 -> lib/
drwxr-xr-x 2 samuel users 4.0K Aug 12 22:27 target/usr/lib/
lrwxrwxrwx 1 samuel users 3 Aug 12 22:27 target/usr/lib64 -> lib/
$ find . -type l -xtype l # find broken symlinks
find: `./host/usr/x86_64-buildroot-linux-gnu/sysroot/lib64/lib': Too many levels of symbolic links
find: `./host/usr/x86_64-buildroot-linux-gnu/sysroot/usr/lib64/lib': Too many levels of symbolic links
./host/usr/x86_64-buildroot-linux-gnu/sysroot/usr/lib/libnss_files.so
./host/usr/x86_64-buildroot-linux-gnu/sysroot/usr/lib/libnss_nis.so
./host/usr/x86_64-buildroot-linux-gnu/sysroot/usr/lib/libnss_compat.so
./host/usr/x86_64-buildroot-linux-gnu/sysroot/usr/lib/libnss_nisplus.so
./host/usr/x86_64-buildroot-linux-gnu/sysroot/usr/lib/libutil.so
./host/usr/x86_64-buildroot-linux-gnu/sysroot/usr/lib/libthread_db.so
./host/usr/x86_64-buildroot-linux-gnu/sysroot/usr/lib/libcidn.so
./host/usr/x86_64-buildroot-linux-gnu/sysroot/usr/lib/libcrypt.so
./host/usr/x86_64-buildroot-linux-gnu/sysroot/usr/lib/libm.so
./host/usr/x86_64-buildroot-linux-gnu/sysroot/usr/lib/libnss_hesiod.so
./host/usr/x86_64-buildroot-linux-gnu/sysroot/usr/lib/libnsl.so
./host/usr/x86_64-buildroot-linux-gnu/sysroot/usr/lib/librt.so
./host/usr/x86_64-buildroot-linux-gnu/sysroot/usr/lib/libnss_db.so
./host/usr/x86_64-buildroot-linux-gnu/sysroot/usr/lib/libanl.so
./host/usr/x86_64-buildroot-linux-gnu/sysroot/usr/lib/libBrokenLocale.so
./host/usr/x86_64-buildroot-linux-gnu/sysroot/usr/lib/libnss_dns.so
./host/usr/x86_64-buildroot-linux-gnu/sysroot/usr/lib/libresolv.so
./host/usr/x86_64-buildroot-linux-gnu/sysroot/usr/lib/libdl.so
./target/etc/resolv.conf
./target/dev/log
After this patch:
$ ls -ld host/usr/x86_64-buildroot-linux-gnu/sysroot/{,usr/}lib* target/{usr/,}lib*
drwxr-xr-x 2 samuel users 4.0K Oct 30 2012 host/usr/x86_64-buildroot-linux-gnu/sysroot/lib/
lrwxrwxrwx 1 samuel users 5 Aug 12 22:36 host/usr/x86_64-buildroot-linux-gnu/sysroot/lib64 -> lib/
drwxr-xr-x 5 samuel users 4.0K Oct 30 2012 host/usr/x86_64-buildroot-linux-gnu/sysroot/usr/lib/
lrwxrwxrwx 1 samuel users 5 Aug 12 22:36 host/usr/x86_64-buildroot-linux-gnu/sysroot/usr/lib64 -> lib/
drwxr-xr-x 3 samuel users 4.0K Oct 30 2012 host/usr/x86_64-buildroot-linux-gnu/sysroot/usr/libexec/
drwxr-xr-x 2 samuel users 4.0K Aug 12 22:36 target/lib/
lrwxrwxrwx 1 samuel users 3 Aug 12 22:36 target/lib64 -> lib/
drwxr-xr-x 2 samuel users 4.0K Aug 12 22:36 target/usr/lib/
lrwxrwxrwx 1 samuel users 3 Aug 12 22:36 target/usr/lib64 -> lib/
$ find . -type l -xtype l # find broken symlinks
./target/etc/resolv.conf
./target/dev/log
Fixes http://autobuild.buildroot.net/results/23fb6b1479d2b5906b72c9437b06ab4700ff246d/
Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Commit fde2605765 (Makefile: test before search for kernel modules)
changed the way we strip kernel modules, but it fails when modules aren't
available (as test -d returns with a non zero exit code).
Fix it by including the test -d call in a proper shell conditional.
Reported-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
The 'find $(TARGET_DIR)/lib/modules' used to find and strip kernel
modules fails when no kernel modules have been installed. While the
'|| true' prevents the entire build from failing, there are still some
error messages displayed, which is not nice.
Instead, test if the directory exists before doing the find. We also
remove the '|| true' in order to really abort the build if a
problematic error occurs.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Luca Ceresoli <luca@lucaceresoli.net>
'cut' needs to be invoked with the -s option to make sure it doesn't
print anything when the delimiter isn't found. This is particularly
important for the charmap detection, because UTF-8 is appended if
the charmap is empty. But without -s, it will never be empty.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This patch moves the host-ccache build target from BASE_TARGETS in Makefile
to an actual host prerequisite in support/dependencies. This causes
host-ccache to be built as part of the dependencies, before any real package
is built.
Since the dependencies are built without ccache anyway, there is no need to
set HOST_CCACHE_CONF_ENV anymore.
Suggested-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Acked-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Add a new goal that dumps all of our variables.
Taken from http://www.cmcrossroads.com/ask-mr-make/6521-dumping-every-makefile-variable
Signed-off-by: Emeric Vigier <emeric.vigier@savoirfairelinux.com>
[yann.morin.1998@free.fr: drop per-variable print; refresh patch]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The -perm +mode is deprecated, and sometimes yields suprising results.
It can be confused with permission in symbolic mode, for example '+u+g',
as POSIX spec suggests.
Signed-off-by: Tzu-Jung Lee <tjlee@ambarella.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
On Gentoo systems with ruby installed RUBYOPT is set and causes build
breakage for some packages that needs host-ruby (example: webkit).
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
It can be useful to have different configuration use the same post-build
and/or post-image scripts as they share a common infrastructure, but yet
have minor differentiation.
This option allows passing zero or more additional arguments to each
post-build or post-image script.
The same set of extra arguments are passed to all scripts, it is not
possible to pass different arguments to each script.
[Peter: fix help text, post-image gets called with the images dir]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Currently, when we need to do a conditional on the type of C library
used, we need to take into account the three toolchain backends. As we
are going to add eglibc support to the Buildroot toolchain backend, it
would become even uglier, so this patch introduces two new hidden
options: BR2_TOOLCHAIN_USES_UCLIBC and BR2_TOOLCHAIN_USES_GLIBC, that
exist regardless of the toolchain backend. The entire Buildroot code
base is converted to use those options.
Note that we have intentionally created only one option
(BR2_TOOLCHAIN_USES_GLIBC) for both glibc and eglibc, since they are
essentially the same, as far as Buildroot is concerned.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
This finally removes the BR2_HAVE_DEVFILES option, that was used to
install/keep development files on target. With the recent migration of
the internal backend to the package infrastructure, we had anyway lost
the ability to build gcc for the target, and install the uClibc
development files on the target.
[Peter: also remove support/scripts/copy.sh]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Instead of using the 'uclibc' target for all toolchain backends,
introduce more sensible target names for the external toolchain and
Crosstool-NG toolchain backend make targets.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
The $(O)/toolchain directory, also called $(TOOLCHAIN_DIR) is no
longer needed, as all packages are now built in $(O)/build/, including
gcc and uClibc.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Commit 14f48861 ("arc: Add ARC and ARC BE architecture") introduced a
sed match against arc* buildroot architecture to translate it to arc.
This causes a problem with armv8, which is called in buildroot aarch64,
which is also matching, resulting in a build failure.
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Also make sure that older kernels are not selected for ARC.
Signed-off-by: Mischa Jonker <mjonker@synopsys.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Initially, a "cp" was used for the one-time target skeleton
installation, and an "rsync" was used for the every-time overlay
skeleton installations. Then, the target skeleton install was
changed to also use rsync (1cd3f992fb).
Now there are two rsyncs, but with different filter specs, which
seems odd.
This patch adds --excludes to both invocations of rsync to make them
consistent, and removes the -u option from the initial target skeleton
installation.
Signed-off-by: Danomi Manchego <danomimanchego123@gmail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Makes the code uniform with the post-build and post-image implementation
(which is slightly simpler and, presumably, more efficient).
Signed-off-by: Luca Ceresoli <luca@lucaceresoli.net>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
This also simplifies code by removing the ifneq/endif clauses.
Signed-off-by: Luca Ceresoli <luca@lucaceresoli.net>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
This also simplifies code by removing the ifneq/endif clauses.
[Peter: drop extra space]
Signed-off-by: Luca Ceresoli <luca@lucaceresoli.net>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
olddefconfig is a new target available in kconfig that allows to take
an old .config file, and update it to the latest version of the code,
without being asked for questions: it automatically assumes the
default value for options whose value was not defined.
It for example allows to take a minimal defconfig, copy it as .config,
and do 'make olddefconfig' to get the full .config.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Tested-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
We already use the .root stamp file to remember if we have copied the
skeleton, so we can remove the /bin check.
Likewise, we should always have a TARGET_SKELETON (or explicitly error
out otherwise), so remove that check as well.
Finally, now that we're using rsync to do the copy, we might as well
use its exclude support instead of cleaning up unwanted files afterwards.
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
An example is .init_enable_core, to enable coredumps in busybox
Signed-off-by: Tilman Keskinöz <arved@arved.at>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
