This can be achieved by adding a new target in target/device/, and I
will later propose a mechanism to add external boards and packages, so
that our users not willing to get their target-specific or
package-specific things merged can keep them cleanly separated from
Buildroot tree.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The "project" feature was designed to allow to several projects to be
built inside the same Buildroot source tree and allowing the toolchain
and non-configurable packages to be shared between the different
projects on the same architecture. While being interesting in theory,
this feature adds a level of complexity to Buildroot, both from an
user perspective and from a developer perspective, while one of the
main Buildroot strengh is to be simple. Moreover, this feature is only
seldomly used by our users.
From a user-level perspective, this for example allows to remove the
project_build_ARCH directory, which was very confusing. The
autotools-stamps directory is also removed, since these stamps are
back at their normal location.
Description of the changes involved :
* project/, directory removed
* Makefile
- Don't include project/Makefile.in and project/project.mk anymore
- Grab a copy of the contents of project/Makefile.in at the
location it was imported, but remove the definition related to
PROJECT_BUILD_DIR. The TARGET_DIR is now in
$(BUILD_DIR)/target_dir
- Remove the creation/removal of the $(PROJECT_BUILD_DIR) and
$(PROJECT_BUILD_DIR)/autotools-stamps directories
- Don't make world depends on target-host-info. This target was
defined by project/project.mk to customize /etc/issue,
/etc/hostname and create /etc/br-version depending on the
project definitions. We can of course imagine re-adding such a
feature later.
- Replace PROJECT_BUILD_DIR by BUILD_DIR everywhere
- Remove the update, log and lognr.$(PROJECT) target, they were
specific to the project feature.
* package/Makefile.autotools.in
- Replace PROJECT_BUILD_DIR by BUILD_DIR for the location of the
configure cache
- Move the INSTALL_TARGET and HOOK_POST_INSTALL stamps to the same
directory as the other stamps (i.e, in the package directory).
* package/Makefile.in
- Replace PROJECT_BUILD_DIR by BUILD_DIR for the location of the
configure cache
* package/at/at.mk,
package/busybox/busybox.mk,
package/busybox/initramfs.mk,
package/customize/customize.mk,
package/linux-fusion/linux-fusion.mk,
package/ltp-testsuite/ltp-testsuite.mk,
package/nfs-utils/nfs-utils.mk,
target/cpio/cpioroot.mk,
target/cramfs/cramfs.mk,
target/device/Atmel/DataFlashBoot/DataflashBoot.mk,
target/device/Atmel/Makefile.in,
target/device/Atmel/at91bootstrap/at91bootstrap.mk,
target/device/KwikByte/Makefile.in,
target/ext2/ext2root.mk,
target/initramfs/initramfs.mk,
target/iso9660/iso9660.mk,
target/jffs2/jffs2root.mk,
target/linux/Makefile.in,
target/romfs/romfs.mk,
target/squashfs/squashfsroot.mk,
target/tar/tarroot.mk,
target/ubifs/ubifsroot.mk
- Replace PROJECT_BUILD_DIR by BUILD_DIR
* target/device/Config.in
- Do not include project/Config.in anymore
* target/linux/Makefile.in.advanced
- Replace PROJECT_BUILD_DIR by BUILD_DIR
- Store the stamps file in $(STAMP_DIR) instead of
$(PROJECT_BUILD_DIR)/autotools-stamps
* target/u-boot/Makefile.in
- Replace PROJECT_BUILD_DIR by BUILD_DIR
- Remove $(PROJECT) from the U-Boot target binary name
- Remove the insertion in the configuration of the project name as
the hostname
- The u-boot-autoscript target now generates
$(U_BOOT_AUTOSCRIPT).img instead of
$(U_BOOT_AUTOSCRIPT).$(PROJECT)
* toolchain/gcc/gcc-uclibc-3.x.mk
toolchain/gcc/gcc-uclibc-4.x.mk
- Move the stamps files to $(STAMP_DIR)
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Define a qstrip function that strips double quotes and then
whitespaces around a given string. This is a very commonly needed
thing, and having a function allows to remove a large quantity of
hand-written definitions, followed by a dummy comment to make text
editors not crazy because of the unmatching double quote.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The find command syntax is incorrect when deleting CVS and .svn directories
from the target filesystem, which prevent CVS directories from being
deleted.
It's useful to delete these dirs because users of buildroot often import
buildroot into CVS or Subversion.
Signed-off-by: Will Newton <will.newton@gmail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
It involved:
* Modifying the top-level Makefile to create the xconfig target, and
the $(CONFIG)/qconf target to compile the qconf utility
* Modifying the package/config/Makefile magic so that C++ files and
binaries gets linked properly, re-using the
package/config/Makefile.kconfig logic inherited from the kernel
* Hacking a little bit package/config/Makefile.kconfig in a way
similar to the modifications done for the mconf (menuconfig
interface)
Signed-off-by: Alper Yildirim <ayildirim@aselsan.com>
The config help text says it all:
Specify a script to be run after the build has finished and before
the BR2 starts packing the files into selected packages.
This gives users the oportunity to do board-specific cleanups,
add-ons and the like, so the generated files can be used directly
without further processing.
The script is called with the target directory name as first and
only argument. Make sure the exit code of that script is 0,
otherwise make will stop after calling it.
Signed-off-by: Daniel Mack <daniel@caiaq.de>
Cc: Sven Neumann <s.neumann@raumfeld.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Remove toolchain build dir, and don't try to remove the non-existing include
dir. Also don't explicitly remove the kernel image located in binaries as
we're already removing that dir.
Add BR2_ENABLE_LOCALE_PURGE / BR2_ENABLE_LOCALE_WHITELIST options to
remove unwanted locales from the target rootfs. Handy for stuff like
the gtk stack, which comes with ~25 MB locales.
Works similar to localepurge in Debian, E.G. you provide a white list
of wanted locales, and everything else is removed.
Move stamp (dependency) files outside the (version specific) source
directories, so other packages can hardcode dependencies on them instead
of having to use <PACKAGE>_VERSION variables.
This is important as the variables in the make rules are evaluated when
the rules is seen, which might be before the dependent makefile is parsed
(and hence <PACKAGE>_VERSION variable is known, screwing up stuff.
The downside of this is that the package isn't automatically rebuilt
when the version changes (E.G. by a svn update) and you now also have to
remove the stamp files next to $(BUILD_DIR)/<PACKAGE>-* to force a rebuild.
I often find myself forgetting what those targets were called, and I imagine
others may not even be aware of their existence.
Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com>
Signed-off-by: Hans-Christian Egtvedt <hans-christian.egtvedt@atmel.com>
This patch will add a rule to top level Makefile to depend on the
$(PROJECT_BUILD_DIR)/autotools-stamps as a required directory. Hence it will be
generated if missing in stead of made when the $(PROJECT_BUILD_DIR)/.root rule
is triggered.
Signed-off-by: Hans-Christian Egtvedt <hans-christian.egtvedt@atmel.com>
This patch will create the autotools-stamps directory early in the build
process, thus making it possible for non Makefile.autotools.in packages to use
this directory to hold stamp files.
Signed-off-by: Hans-Christian Egtvedt <hans-christian.egtvedt@atmel.com>
HOST_GLIB is set to the path that contains the host glib tool set and is
used when building packages using glib. The buildroot top level Makefile
sets HOST_GLIB using which to find the path where glib-genmarshal is
located.
The problem is that a cross compiled version of glib-genmarshal is also
put in the build_ARCH/staging_dir/usr/bin directory when the package
libglib2 is built. This cross compiled version will typically not run on
the host system.
Fix it by ignoring staging_dir in the which output.
Closes#5934
jacmet: fixed to work correctly if it's only found in staging_dir.
Improve external toolchain support
* Do not put kernel-headers in the dependencies of BASE_TARGETS in
the case where BR2_TOOLCHAIN_SOURCE is not y. The kernel headers
are already supposed to be part of the external toolchain, so
there's no need to download, extract and install them.
* In the configuration system, don't display the kernel headers
version selection list when an external toolchain is selected. This
is implemented by moving the source
"toolchain/kernel-headers/Config.in" inside the if
BR2_TOOLCHAIN_SOURCE in toolchain/Config.in.2.
* Change the description and help message of the BR2_LARGEFILE,
BR2_INET_IPV6, BR2_INET_RPC, and BR2_SOFT_FLOAT option in
toolchain/external-toolchain/Config.in. In the case of an external
toolchain, the semantic of these options is not to enable large
file support, IPV6 or RPC (since the toolchain is already compiled,
it has been decided previously). Their semantic is to let Buildroot
know about the characteristics of the external toolchain being
used.
As an improvement, we could guess these values automatically:
- for BR2_LARGEFILE, look at the value of __UCLIBC_HAS_LFS__ in
bits/uClibc_config.h in the libc headers directory.
- for BR2_INET_RPC, look at the value of __UCLIBC_HAS_RPC__ in the
same file
- for BR2_INET_IPV6, look at the value of __UCLIBC_HAS_IPV6__ in
the same file
- for BR2_SOFT_FLOAT, look at the output of $(CC) -v 2>&1 | grep
-- "--with-float=soft"
But I'm not sure how this would be possible, since these values are
used at configuration-time by other configuration options, not only
at build time.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
make "make clean" remove the .root stamp, so that everything will
be reinstalled properly on a subsequent build. (bug id 4304)
Also remove new autotools-stamps directory.
The MAKELEVEL test is not that robust; It fails with the recent log
support or if buildroot is driven from an external Makefile.
Rework it to instead detect source-check/external-deps by the fact that
they set SPIDER.
Mark empty directories in target skeletons with .empty files
and remove them while copying to target (like it's already done for
svn and cvs files) to better support version control system which don't
handle empty directories (E.G. git-svn).
Will need to try to copy eventual pre-existing project-specific deps back
to package/config in order not to mess up the corresponding timestamps (to avoid superfluous rebuilds)..
=========================================================
The purpose of the BSP patch is to allow building
several boards inside the same buildroot tree.
For this to work, each board has to have its
own "$(TARGET_DIR)" and all *configurable* packages
must be rebuilt for each board.
They are now built in the "$(PROJECT_BUILD_DIR)"
All non configurable packages can and should still
be built in the "$(BUILD_DIR)".
If a package is built for one board, then when
you build for a second board of the same architecture
the build becomes a simple copy of the resulting
binaries.
-----
Define BR2_PROJECT which will be used as the selector
between different boards. Note that BR2_PROJECT allow
you to build multiple root file systems for a single
board, and should not be confused with BR2_BOARD_NAME
which relates to the H/W.
-----
Define PROJECT_BUILD_DIR as "PROJECT_BUILD_DIR/$(PROJECT)"
Define BINARIES_DIR as "binaries/$(PROJECT)"
Define TARGET_DIR as "$(PROJECT_BUILD_DIR)/root"
(some prefix/postfix may apply)
Resulting images are stored in "$(BINARIES_DIR)"
-----
Define a few new environment variables in Makefile
PROJECT: Stripped BR2_PROJECT
DATE: Date of build in YYYY-MM-DD format
HOSTNAME: Stripped BR2_HOSTNAME => /etc/hostname
BANNER: Stripped BR2_BANNER => /etc/issue
Linux and Busybox will be built in $(PROJECT_BUILD_DIR)
More patches will be needed later to ensure all
configurable packages are built in this directory.
This also provides a nice up-to-date filestamp if something got reconfigured (e.g. c++ support got toggled) that can be used in the future to depend accordingly.
This patch changes the way the top level Makefile searches for a
<board>_defconfig file, it will only look in the target/ directory and its sub
directories.
The patch also enables loading a defconfig even if there already is a .config.
(Hans-Christian Egtvedt)
wget's spider-mode (i.e. just see if the server says the file is there without
actually downloading the files.
Convenient to check for unavailable source-packages.
- depending on the BR2_GNU_BUILD_SUFFIX, set the respective EXEEXT, LIBEXT, SHREXT extensions for use on the target.
Thanks to Tom for suplying a diff which implements these.
Quick patch to avoid deleting my directory of RPM Sources.
I tend to use buildroot as "make DL_DIR=/.rh/SOURCES", with the rpm
topdir set to $HOME/.rh in my .rpmmacros -- allows me to build RPMs as
non-root, and share the source with buildroot. Problem is, buildroot
likes to delete my directory :(
and Dan Kegel's crosstools (7??-*).
Use makefile vars for toolchain names in ccache.mk and gdb.mk.
Fix a g++-related symlink issue with ccache.
Change the default to gcc 3.4.2 and binutils 2.15.91.0.2 (now that mips
is building). Let the users go forth and test... ;-)
1) Allow selection of binutils/gcc/gdb/kernel headers to build, although
some of the older tool patches probably need updating.
2) Rework gdb build so that remote debugging now works with gdbserver.
3) Misc. other package updates.3) Misc. other package updates.3) Misc. other package updates.
causing problems. Newer uml messes up sigcontext.h and is a pain
to build anyways. So kill it for now. It is trivial to build for
stock i386 and then boot the generated root_fs using uml.
works for dynamicly linked executables. Only tested so far on i386 with
gcc 3.3.2. But combined with the latest ldso updates, I can now get a
dynamicly linked java helloworld app to run.
Add a config toggle for building gcj and libgcj. The build runs fine
through gcc-final but currently requires hand interventions to build
gcc-target. Again, only tested so far on i386.
issues for applications, we also create staging_dir/bin/<arch>-linux-*
symlinks which allows us to configure most apps with target=<arch>-linux.
gcc 3.3.2 libstdc++ now correctly identifies uclibc locale support.
There are still some issues (mainly related to wide char time strings
and wcsftime) to be addressed, but those are on hold until the uClibc
locale internals are reworked once again.
With the new stuff, we can also build gcc 2.95 and STLport again, although
the native gcc build for the target is failing.
Archs supported (some archs not yet tested):
2.95 - i386, arm, mips*, powerpc.
3.3.2 - Hopefull all for which uClibc has shared lib support.
i386 (limited to 64 bit long double... same as double), arm (libfloat),
mips, and mipsel.
Enable cross compiling a native gcc 3.3 toolchain to run on the target.
Misc rootfs cleanups... strip some things that weren't, eliminate duplicate
libs, move openssl shared libs out of /lib, reduce size of libssl.so
by dynamicly linking with libcrypto.so, fix dropbear compile on mips.
Add support for libstd++ in full locale mode, and remove the need for
config hacks when using uClibc stub locale support.
TODO --
gdb intl support is still broken.
uClibc should auto-download the pregenerated locale data
Thomas Petazzoni