This commit fixes two separate, but related build failures:
* A failure that was happening when the ARM CPU supports NEON,
floating point is enabled, but not with the NEON FPU. In this case,
the NEON ARM assembly is rejected by the assembler, with messages
like "Error: selected FPU does not support instruction -- `vmul.f32
q0,q0,q1'". To fix this, we pass -mfpu=neon when we build mplayer
with NEON support.
Fixes:
http://autobuild.buildroot.org/results/257/257a10e9cb5022bb09e0c6a03844be5b5b3e0bd4/
* A failure that was happening when the ARM CPU supports NEON, but
the configuration is anyway using soft-float. In this case, mplayer
attempts to compile NEON floating point instructions, but this
obviously fail in a soft-float context, with errors such as 'Error:
selected processor does not support ARM mode `vmov d0,r7,r8''. To
fix this, we do not allow NEON to be enabled when we are in a
soft-float configuration.
Fixes:
http://autobuild.buildroot.org/results/7b3/7b3c89fcd496c0bc80063f63ecd58c827e8077ea/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
In most embedded systems, there is not much point in installing bash
completion files, especially since bash may not necessarily be present
on the target.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
E2tools is a simple set of GPL'ed utilities to read, write, and
manipulate files in an ext2/ext3 filesystem. These utilities access a
filesystem directly using the ext2fs library.
[Thomas: add toolchain dependencies as needed, use full Git hash, use
github helper, add support for host version.]
Signed-off-by: Daniel Nyström <daniel.nystrom@timeterminal.se>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
vo-aacenc has some hand-written assembly that cannot build in Thumb2,
so we have to explicitly pass -marm to make sure this particular
package is always built with full-length ARM instructions.
Fixes:
http://autobuild.buildroot.org/results/680/680b29cd824624eb8e4ec71187b9a6576444e72b/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This reverts commit ea737fb100, which in
the end didn't fix the autobuilder problem. Since then,
fb80d28341 ('nodejs: add a patch forcing
link command to use CXX') was added, and was successfully tested.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Until now, our UPPERCASE function was implemented purely in make for
performance reasons, but our LOWERCASE function was implemented by
calling "tr", which was reasonable due to the fact that LOWERCASE was
rarely used, but future changes might make a more heavy usage of the
LOWERCASE macro.
We want this LOWERCASE function to turn a "_" into a "-" and not a
".", so we slightly adjust the existing FROM and TO lists to make this
possible. This doesn't change the behavior of the UPPERCASE macro
because both "-" and "." are converted into "_" by this function.
This change takes advantage of suggestions made by Arnout
Vandecappelle, who said they further improve the performance of
UPPERCASE and LOWERCASE by 30%/40%.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Based on suggested new package by Marco Trapanese ([1]).
[1] http://lists.busybox.net/pipermail/buildroot/2014-February/090661.html
[Thomas: make it only available with glibc toolchains.]
Signed-off-by: Peter Seiderer <ps.report@gmx.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The cross-compilation improvements integrated in Python rely on the
compiler exposing a line starting with LIBRARY_PATH when called with
-E -v. This is used by Python setup.py to find the installation
locations of libraries.
However, this LIBRARY_PATH line is not shown by very old compilers,
such as the gcc 4.2.x compiler used on the AVR32 architecture. This
causes libraries installed in the sysroot, such as libffi, to not be
detected by the setup.py script.
To fix this problem, this patch adds addtional logic to setup.py,
which consists in deriving the library paths from the sysroot
location, if no LIBRARY_PATH field was found.
Fixes:
http://autobuild.buildroot.org/results/7a6/7a65e381cc04bf8f74fd63a6dcda502f3c26aeef/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The cross-compilation improvements integrated in Python rely on the
compiler exposing a line starting with LIBRARY_PATH when called with
-E -v. This is used by Python setup.py to find the installation
locations of libraries.
However, this LIBRARY_PATH line is not shown by very old compilers,
such as the gcc 4.2.x compiler used on the AVR32 architecture. This
causes libraries installed in the sysroot, such as libffi, to not be
detected by the setup.py script.
To fix this problem, this patch adds addtional logic to setup.py,
which consists in deriving the library paths from the sysroot
location, if no LIBRARY_PATH field was found.
Fixes:
http://autobuild.buildroot.org/results/1aa/1aad2d677dbf7b0a3cb0120a35f45123124f51ab/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
[Thomas: remove BR2_TARGET_ROOTFS_YAFFS_PATH option, rename to yaffs2
instead of yaffs.]
Signed-off-by: Will Wagner <willw@carallon.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
[Thomas: remove target variant of the package, use labelled version
0.2.9 instead of latest.tar.gz, use 'make install' instead of doing
manual installation, add license informations.]
Signed-off-by: Will Wagner <willw@carallon.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Add an option for embedding a config file directly in grub.
Signed-off-by: Dima Zavin <dmitriyz@google.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The target depends on the toolchain so add it as a dependency.
This also fix the support to top-level parallel make.
Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The "toolchain-eclipse-register" target needs the toolchain so add
toolchain as a dependency.
This also fix the support to top-level parallel make.
Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The host variant of syslinux wasn't exposed anywhere, so the only way
to use it was to know that a "host-syslinux" target existed. Moreover,
thanks to commit 8e0d411898 ('syslinux:
install helper programs to the host'), the host utilities of syslinux
are automatically installed to $(HOST_DIR) when building the target
syslinux.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Even though syslinux is built for the target, it does create helper
programs intended to be run on the host such as syslinux and extlinux.
This change installs these helper programs to the host so that they may be
called by post image creation scripts.
Signed-off-by: Frank Hunleth <fhunleth@troodon-software.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The current check is broken, as it only checks if the user's login
shell is bash, not what the system shell is.
Mimick the sequence found in the top-level Makefile to search for
bash, except for the fallback case, where we explicitly check that
'sh' is bash, by checking if it sets $BASH, so we know the fallback
case, in the top-level Makefile, to use 'sh' will indeed use bash.
Remove superfluous semi-colons ';' at the end of lines, they are
not needed in a shell script (this is not C!)
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Peter Korsgaard <jacmet@uclibc.org>
Cc: Andrew Barnes <andy@outsideglobe.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
In commit 971e331c54
('package/linux-headers: remove 2.6 snapshot'), we removed the support
for the Linux headers snapshot option, but we forgot to remove some
related help text in the main Config.in option of the Linux headers
package.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
libjansson.so doesn't list libm in its NEEDED header. Fixes the following
nftables config time failure:
.../usr/lib/libjansson.so.4: undefined reference to `__isnan'
.../usr/lib/libjansson.so.4: undefined reference to `__isinf'
Fixes:
http://autobuild.buildroot.net/results/152/1525d7a3e1d1fcf35858962251c0b69a5e1b64db/
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Strictly speaking, linuxthreads (new) should be enough, since it includes
pthread_barrier_* implementation. However, we currently have no clean way to
distinguish between LT and LT.old.
Fixes:
http://autobuild.buildroot.net/results/66a/66a9d2c664e0a9b2b0c51b998c72c2ceda90ff9b/
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Currently, we only pass --enable-opengl is at least one DRI driver is
enabled, since full OpenGL is only possible with a DRI driver (and the
X.Org stack). Otherwise, we pass --disable-opengl on the assumption that
OpenGL is not possible.
But mesa3d's ./configure is a bit weird: enabling OpenGL-ES and diabling
full OpenGL is an error (see autobuild failure below).
It turns out that, if there is no DRI driver enabled, but --enable-opengl
is passed, ./configure will behave properly wrt OpenGL-ES, and will not
build the full OpenGL libgl.
Thanks to Paul for explaining this.
Fixes:
http://autobuild.buildroot.net/results/61c/61c1c566dc0e829cb663ca30b1fd6cf9cc6cd931/
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Paul Cercueil <paul@crapouillou.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Currently, graph-depends (and PKG-graph-depends) do not store the
intermediate 'dot' program.
Some users would like to get the dot program to be able to further
customise the generated graphs (eg. modify the layout, colorise some
of the packages...)
So, store the intermediate dot program alongside the generated graph.
Reported-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Acked-by: "Maxime Hadjinlian" <maxime.hadjinlian@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The Lua binding option of libuci doesn't need the MMU.
Signed-off-by: Hadrien Boutteville <hadrien.boutteville@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
