The host-kmod version of depmod gets installed into HOST_DIR/sbin, so
adjust the path we use to refer to it in linux.mk
Reported-by: Ryan Coe <bluemrp9@gmail.com>
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>
Specifying a floating tag like HEAD for a repository version is bad practice,
as it results in non-reproducible builds. This patch removes the default
assignment of HEAD as version when a custom git repository is used for the
Linux kernel.
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The BR2_LINUX_KERNEL_INSTALL_TARGET option not only copies the kernel
image to /boot, but also the DTBs, if such DTBs are generated by the
build process. It makes sense to mention that in the help text of this
option.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
When installing DTB to the images directory, do this as part of the image
install commands rather than the target install commands.
Signed-off-by: Reuben Dowle <reuben.dowle@4rf.com.az>
Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The target for building the kernel with a cpio initramfs is not
calling the append dtb commands, creating a final kernel image
without an appended dtb. Instead it needs to call kernel make on
the intermediate target, then call the append dtb commands
to produce the final image as the primary kernel build target
does.
Signed-off-by: Nathan Ford <nford@westpond.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Fixes bug #5516 - appended device tree blobs on uImage fails
Before version 3.7 of the kernel, building the zImage and then the
uImage will rewrite the zImage in the process, removing the device tree
we just appended.
Use mkimage to append the device tree to the uImage and rebuild the
headers directly.
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
We're only EABI now, so we want to always build an EABI kernel
when we're building for ARM.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Initramfs compression does not make much sense for the architectures
that support compressed kernel images because in this case the data
would be compressed twice. This will eventually result in a bigger
kernel image and time overhead when uncompressing it.
The only reason to use compressed initramfs is to reduce memory
usage when the kernel prepares rootfs, and both the unpacked
filesystem and initramfs.cpio are present in the memory.
Buildroot attempts to force GZIP compression for initramfs,
however it doesn't always work because initramfs compression mode
depends on RAM disk compression supported by the kernel.
Thus, CONFIG_INITRAMFS_COMPRESSION_GZIP depends on CONFIG_RD_GZIP.
If CONFIG_RD_GZIP is not set, setting GZIP initramfs compression
will have no effect.
Besides, the kernel also supports other compression methods,
like BZIP2, LZMA, XZ and LZO. Forcing the good old GZIP does not
really make much sense any more.
This removes initramfs compression settings from Buildroot,
so that the default value preset in the kernel config is used,
which is CONFIG_INITRAMFS_COMPRESSION_NONE.
If initramfs compression is still needed, it can be set
in the kernel config (using make linux-menuconfig)
Signed-off-by: Valentine Barshak <gvaxon@gmail.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
When the LZO compression mode is chosen the linux kernel requires
lzop host utility at build time.
Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Prefer xz compressed tarball so some bandwidth is saved for kernel headers
and kernel itself downloads.
Signed-off-by: Raúl Sánchez Siles <rasasi78@gmail.com>
Reviewed-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Use a option name that doesn't change for every new 3.x version.
It's easier to maintain and consistent with barebox.
Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Remove the OCF linux kernel extension instead opting to build ocf-linux
modules out of tree.
This is easier for users since no kernel config tweaking is required.
On the downside the OCF drivers can't be used, but then all of the
kernel crypto drivers are available to users via cryptosoft which is
preferred.
Also remove it from the menu to utilize a virtual.
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Commit c461e64201 bumped the linux 3.8.x
stable version from 3.8.5 to 3.8.6, but did so in only one of the two places
in which the version number is hard-coded in linux/Config.in
Signed-off-by: Simon Dawson <spdawson@gmail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
The option BR2_LINUX_KERNEL_INSTALL_TARGET used to install the kernel in
/boot in the rootfs. The introduction of the device tree builds made it
possible to generate device tree blobs, and stored them in IMAGES_DIR
like the kernel, but didn't copy the dtb in /boot when _INSTALL_TARGET
was enabled.
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Closes#5846
The $(dir ...) function leaves a slash at the end, so that the LINUX_SITE
variable for a custom tarball ends in a slash. The DOWNLOAD macro adds
another slash between SITE and SOURCE, which results in a double slash in
the download URL.
Fix this by stripping off the final slash from the _SITE in all packages that
have a custom tarball.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
The $(wildcard ) doesn't work for LINUX_APPEND_DTB, because the .dtb
doesn't exist yet at that point.
Also factor the common part out.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
.dtb files are now generated in arch/$ARCH/boot/dts instead of
arch/$ARCH/boot, so extend the LINUX_INSTALL_DTB rule to look there
as well.
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Only qstrip once and add a KERNEL_DTBS helper variable to simplify the dtb
rules.
No functional change.
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Newer kernels had a slight change in the random number API so the OCF
patchset requires a slight tweak to build and run properly.
Affected kernels are, per series:
>=3.0.41, >=3.2.29, >=3.4.9, >=3.5.2 and any newer version.
The fix is applied by detecting the API change rather than guessing the
kernel version.
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
The Xtensa architecture had been removed because it required special
handling and depended on additional directories and files that became
obsolete over time. This change is more aligned to other architectures.
[Thomas: rebased on top of the "arch: improve definition of gcc mtune,
mcpu, etc." patch].
Signed-off-by: Chris Zankel <chris@zankel.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
As stated in commit 555c2585bf, the
Xtensa architecture has been introduced in 2009 and never changed
since its initial introduction. It requires some special handling that
is a bit annoying, and despite our call to the initial developers, and
the announcement of the deprecation of the architecture during the
2012.05, nothing has happened. Therefore, drop support for this
architecture.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: me
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Consist with other architectures.
Signed-off-by: Sonic Zhang <sonic.zhang@analog.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This patch adds support for the ARM-only appended device tree
mechanism present in the kernel.
This option allows to add at the end of the kernel image the
device tree blob so that we can still boot device tree enabled
kernels with old bootloaders.
This patch also adds the needed logic to genereate such an image
when building zImages or uImages, also adding the necessary parts
to rebuild the uImage.
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This patch introduces the BR2_LINUX_KERNEL_UBOOT_IMAGE boolean to
factorize more code that will be shared in the next patches that
introduces other uImage-like targets.
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This patch introduces some support for device tree-enabled kernels.
It replaces the former BR2_LINUX_KERNEL_DTS_FILE option that was
microblaze-only, that was quite limited. This option was quite
limited, first obviously because it was restricted to microblaze,
but also because it targetted only external device tree source files,
and allowed only to build simpleImages using the custom image name
mechanism.
This patch adds a much more generic one, that can work on basically
every architecture that supports device tree. It allows to build
both device tree source file that comes with the kernel source or to
set the path to the device tree file to use so that one can use a
custom device tree.
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This is way too specific to microblaze-only. Remove this support to
introduce a more generic way to do support device tree kernels.
This reverts commit aaed42d156.
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The help text of BR2_LINUX_KERNEL_DEFCONFIG is confusing: it seems to
point to the defconfigs in the buildroot tree. So explicitly refer to
the kernel tree in the help text.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Having dtc as a host tool can be useful for users that have a custom
boot scenario where the device tree is not embedded in the kernel.
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>
Add the ocf-linux package and linux kernel extension.
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
This patch improves two things:
* It makes sure that the linux-menuconfig and al. commands can be
executed even if the user hasn't executed 'make' before. Until now,
the commands were depending on the .configured stamp, which is not
sufficient since the linux package has been converted to the
GENTARGETS infrastructure. Instead, depend on the linux-configure
phony target.
* It makes sure that those commands are not available when the Linux
package is not selected, just like we do for Barebox.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
test should exit with Exit-Code 0 if no .ub-File present and copy the
file if Exit-Code 1, otherwise make fails
Signed-off-by: Markus Kaindl <markus.kaindl@stusta.mhn.de>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
This patch adds the options needed to build the SimpleImage containing
the device tree structure needed for the Microblaze architecture.
Handling Device Tree and SimpleImage will be handled in a general way
in the future.
I provide this patch to be able to build the system in the meantime.
Signed-off-by: Stephan Hoffmann <sho@relinux.de>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
This modifies the definition of DOWNLOAD to receive two arguments:
the first one is the full URL of the file to download, whereas the second
(and optional) is the name the file will have once downloaded.
Same thing with the SOURCE_CHECK_WGET and SCP functions.
All calls to these functions have been changed to the shortest form of
the new API, except for toolchains acquisition. Since there is quite a
number of different toolchains this call to DOWNLOAD is better set to the
generic one.
Signed-off-by: Alvaro G. M <alvaro.gamez@hazent.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Tested-by: Stephan Hoffmann <sho@relinux.de>
Downloading Microblaze LE toolchain works on a clean install
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Drop the patch for kernel >= 3.3
Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
There already is a default kernel version, so it makes little sense to have
a default custom kernel version as well. This default breaks the 'make
savedefconfig' if we want to force a specific version.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
The 'vmlinux' target in the linux build directory is a generic target. It may
not be directly bootable for all architectures, but for projects where a custom
bootloader is used, it can be of value.
Previously, this target was only available for mips architectures. This patch
removes this restriction.
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>
In analogy to build targets uclibc-update and busybox-update, add
extra targets to copy the current configuration to the custom
configuration file set in buildroot. These targets facilitate the
work of developers adding support for a specific board.
linux-update-config copies .config directly.
linux-update-defconfig first creates a defconfig and copies that.
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>
Signed-off-by: Kelvin Cheung <keguang.zhang@gmail.com>
Acked-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
The vmlinux kernel target is valid for powerpc kernels too.
In fact this broke the qemu sample config.
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Bump default kernel vesion to 3.1 to match headers.
Also implement downloads for 3.x series kernels.
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Thanks to the pkgparentdir and pkgname functions, we can rewrite the
GENTARGETS macro in a way that avoids the need for each package to
repeat its name and the directory in which it is present.
[Peter: pkgdir->pkgparentdir]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
An initramfs is in fact the same as a cpio archive, but embedded in
the kernel. So instead of duplicating the cpio infrastructure,
we can simply build images/rootfs.cpio and link that into the
kernel.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
The post extract hook point is not really correct as what RTAI and
Xenomai extensions are doing is patching the kernel.
The post patch hook point doesn't work, because RTAI and Xenomai
patches would be applied *after* all other patches, while it sounds
more logical to apply them first, and *then* allow the user to apply
some platform/board specific patches if needed.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
[Fixes by Thomas, including comments from Arnout:
* Use AUTOTARGETS instead of GENTARGETS
* Use $(KERNEL_ARCH) instead of $(BR2_ARCH) as argument to the
prepare-kernel.sh script. This allows the arch name to be fixed
with the usual sed expressions and the quotes to be stripped.
* Add the --verbose option to prepare-kernel.sh. This allows to get
some clear error message when no Xenomai patch has been found for
the current kernel version.
* Improve the help texts as suggested by Arnout, and remove the
now useless README file.
* Add a BR2_PACKAGE_XENOMAI_SMP option, instead of poking inside the
kernel configuration to find out whether SMP is enabled or
not. This cannot work because: 1) the kernel might be built
outside of Buildroot and 2) if the kernel is built inside
Buildroot, it is built *after* Xenomai, so the kernel
configuration file is typically not yet present.
* Simplify the ARM subarchitecture selection as suggested by Arnout.
* Remove the documentation and development files according to
Buildroot standards (using BR2_HAVE_DOCUMENTATION and
BR2_HAVE_DEVFILES).
* Simplify the /etc/ld.so.conf modification logic.
]
[Peter: whitespace fixes]
Signed-off-by: Benoit Mauduit <benoit.mauduit@openwide.fr>
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Add a kernel sub-menu (called "Linux Kernel Extensions"), which makes
possible to patch it or tweak the kernel build step.
* All linux/linux-ext-*.mk files will be read by the make process.
* The menu can be customized in "linux/Config.ext.in".
[Peter: small fixes]
Signed-off-by: Benoit Mauduit <benoit.mauduit@openwide.fr>
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
The name "patch-kernel.sh" is a bit stupid, since this script is used
to patch everything in Buildroot, not only kernel trees.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Move fw_printenv / fw_setenv options from the uboot bootloader build to
the uboot-mkimage package, and rename it to uboot-tools.
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
The assumption that all kernels are in
http://www.kernel.org/pub/linux/kernel/v2.6/ is no longer true:
versions 3.x are in separate directories.
We now compute the directory name from the major and minor versions of
the version provided by the user. This assumes that the 3.1 version
will be in a /v3.1/ directory, which we don't know yet because the 3.1
cycle hasn't started yet.
At the same time, we add support for the official -rcX versions.
Patch tested by compiling 3.0-rc6, which Buildroot has successfully
downloaded and built.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
As the kernel and bootloaders do not use the normal BR2_PACKAGE_*
Kconfig options, their target name was not automatically added to the
global TARGETS variable. Each bootloader .mk and the linux.mk had to
add their own target manually to TARGETS, and the package
infrastructure was making tests on non-existing Kconfig variables.
This commit improves the package infrastructure so that it looks at
BR2_PACKAGE_<pkg> for packages, BR2_TARGET_<pkg> for bootloaders and
at the special BR2_LINUX_KERNEL for the linux package.
This allows to simplify a little bit the bootloaders and linux .mk
files.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
[Peter: always copy to output/images as requested by Luca]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Four methods for the creation of device files in /dev are now
proposed:
- static method uses device table as before
- devtmpfs method enables this feature in kernel
- mdev method adds mdev starting script to the file system
and selects mdev itself for installation
- udev method selects udev for installation
All dynamic methods are based on devtmpfs, so one doesn't need to care
about /dev folder.
Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
This is useful for ARCHes like PowerPC that have a whole lot of
special targets for each different board. The kernel image target
tells make which OpenFirmware machine description file is combined into
the kernel binary.
[Peter: fix long lines]
Signed-Off-By: Martin Hicks <mort@bork.org>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
[Peter: don't allow MMU on bfin]
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Just needed to pass in ccache as a prefix to the CROSS_TARGET variable.
Signed-off-by: Martin Hicks <mort@bork.org>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
As per the discussion at
http://lists.busybox.net/pipermail/buildroot/2010-December/040030.html,
remove the bits that forced the IPv6 configuration in the kernel
depending on the toolchain ability to support (or not) IPv6. You may
have a toolchain with IPv6 support but still don't want to have IPv6
in your kernel.
The only parameters we adjust in the kernel configuration are:
* ARM EABI, since we got a lot of bug reports regarding misconfigured
kernel compared to the ABI used by userspace applications.
* initramfs, since its contents are generated by Buildroot itself
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
linux-% shortcut targets (short for linux26-%) ignores the ouput dir
$(O) so that 'make O=output.arm linux-menuconfig' is actually run in the
default $(O) directory output/ and not in output.arm/. Fix by passing on
$(O) if set.
[Peter: Use EXTRAMAKEARGS]
Signed-off-by: Bjørn Forsman <bjorn.forsman@gmail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Both i386 and x86_64 architectures are supported by the arch/x86
directory in the kernel. So, when we copy the kernel configuration
file to arch/$(KERNEL_ARCH)/configs/, it does not work because
arch/i386 and arch/x86_64 do not exist.
So, we introduce KERNEL_ARCH_PATH, which is the path to the
architecture specific directory in the kernel source tree.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
The ELF vmlinux image found at the root of the kernel source tree is
the format that Qemu needs when emulating mips(el) or ppc targets, so
add support for it.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
The kernel being a component that often needs a fairly important set
of changes to be adapted to a particular hardware platform, having
maximum flexibility on the patching process is a nice
thing. Therefore, as per the discussions from the Buildroot Developer
Day, we add a mechanism to apply a list of patches (that could come
either from URLs, local files or local directories).
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Now that minimal kernel defconfigs are used in Buildroot, the problem
is that copying those minimal configuration files to .config in the
kernel source tree does not work, as kconfig will ask interactively
what should be the value for all unspecified options.
On suggestion on Sam Ravnborg, the easiest way to solve this is to
import the minimal defconfig file as a defconfig inside the kernel
tree (in arch/$(ARCH)/configs) and configure the kernel with it.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
The linux.mk rewrite lost the DEPMOD setting while installing modules
which means depending on host-module-init-tools has been useless.
Instead, the build system has been executing /sbin/depmod.
While we're here, drop the INSTALL_MOD_PATH since LINUX26_MAKE_FLAGS
already contains it.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
We only have one Linux kernel package, and "linux26-" is an anachronism
in today's world. So add useful "linux-%" shortcuts to the "linux26-%".
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Building with linux-2.6.36 and initramfs support causes the build to
pause while it prompts for newer options (uid/gid/compression). So
have the build system inject the newer options into the linux config
automatically. Older versions should just ignore these.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
The current linux code overrides LDFLAGS that the kernel itself might be
setting up. Looking at the history, there doesn't seem to be any reason
for this override. It was added in ea8b1fa6a6 without any logic.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Linux has been using "usr/initramfs_data.cpio" for a few releases as the
generated cpio name, so the buildroot match of "...cpio.*" won't actually
clean out the previous result.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
And all the infrastructure surrounding it. A broken sed implementation
is quite rare nowadays, as seen by the fact that the current host-sed
support has been broken for a while, so just get rid of it.
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
On most architectures, the kernel image can be found in
arch/<ARCH>/boot, but on AVR32, it's in arch/<ARCH>/boot/images.
Issue initially reported by Joachim Pihl
<joachim.pihl@sensordevelopments.com>.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Some Linux kernel output image formats are available on some archs,
some not. For example 'uImage' is not supported on MIPS, so let's
prevent the user from making this selection.
Issue initially reported by Choi, David <David.Choi@Micrel.Com>.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
When initramfs was not selected as a root filesystem, we forcefully
disabled the initramfs in the kernel configuration.
However, it prevents an user from manually managing its initramfs, as
we override the option he has set in his kernel configuration
file. There's no real reason to do so: when initramfs is not selected
as the root filesystem, just don't touch initramfs related options in
the kernel configuration.
Problem reported by Sergey Naumov <sknaumov@gmail.com>.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Just as we do for U-Boot, error out in the Linux kernel makefile when
the defconfig name or the configuration file path are not
correct. What prompted me to implement this was a report on IRC from
an user using BR 2010.05 and not understand why the kernel build
process was failing. It was because he just forgot to set the path of
the configuration file.
Of course, it doesn't catch all mistakes (like pointing to a
non-existing defconfig or to a non-existing configuration file), but
it at least catches basic mistakes.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
In Buildroot, the kernel is built and installed *before* the root
filesystems are built. This allows the root filesystem to correctly
contain the kernel modules that have been installed.
However, in the initramfs case, the root filesystem is part of the
kernel. Therefore, the kernel should be built *after* the root
filesystem (which, in the initramfs case simply builds a text file
listing all files/directories/devices/symlinks that should be part of
the initramfs). However, this isn't possible as the initramfs text
file would lack all kernel modules.
So, the solution choosen here is to keep the normal order: kernel is
built before the root filesystem is generated, and to add a little
quirk to retrigger a kernel compilation after the root filesystem
generation.
To do so, we add a ROOTFS_$(FSTYPE)_POST_TARGETS variable to the
fs/common.mk infrastructure. This allows individual filesystems to set
a target name that we should depend on *after* generating the root
filesystem itself (contrary to normal ROOTFS_$(FSTYPE)_DEPENDENCIES,
on which we depend *before* generating the root filesystem).
The initramfs code in fs/initramfs/initramfs.mk uses this to add a
dependency on 'linux26-rebuild-with-initramfs'.
In linux/linux.mk, we do various things :
* If BR2_TARGET_ROOTFS_INITRAMFS is enabled (i.e if initramfs is
enabled as a root filesystem type), then we create an empty
rootfs.initramfs file (remember that at this point, the root
filesystem hasn't been generated) and we adjust the kernel
configuration to include an initramfs. Of course, in the initial
kernel build, this initramfs will be empty.
* In the linux26-rebuild-with-initramfs target, we retrigger a
compilation of the kernel image, after removing the initramfs in
the kernel sources to make sure it gets properly rebuilt (we've
experienced cases were modifying the rootfs.initramfs file wouldn't
retrigger the generation of the initramfs at the kernel level).
This is fairly quirky, but initramfs really is a special case, so in
one way or another, we need a little quirk to solve its specialness.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
These targets allow the user to customize the configuration of the
Linux kernel. After changing the kernel configuration, the next time
the user runs "make", the kernel is rebuilt to take into account the
new configuration (not rebuilt from scratch).
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
We only adjust the configuration or ARM EABI and IPv6. The (more
complicated) initramfs case is handled in a separate commit. The user
is expected to take care of all other configuration details (like
having Netfilter enabled to make iptables work, etc.)
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
In order to not depend on module init tools being installed on the
development environment of the Buildroot user, let's build module init
tools for the host.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This patch introduces a single, simple, infrastructure to build the
Linux kernel. The configuration is limited to :
* Kernel version: a fixed recent stable version, same as kernel
headers version (for internal toolchains only), custom stable
version, or custom tarball URL
* Kernel patch: either a local file, directory or an URL
* Kernel configuration: either the name of a defconfig or the
location of a custom configuration file
* Kernel image: either uImage, bzImage, zImage or vmlinux.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
