When BR2_ARM_CPU_HAS_NEON=y, we pass --enable-neon. However, when
BR2_ARM_CPU_HAS_NEON is disabled, we don't pass anything. This generally
works fine, but turned out to integrate NEON code in ARMv7-M
builds (since it's ARMv7, ffmpeg assumed it should enable NEON code).
Even though ffmpeg is now disabled for ARMv7-M, it still makes sense to
be explicit, and disable NEON support when the CPU doesn't have it.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
sg3_utils grep for '^linux' as the host OS to decide if we are
targetting Linux.
However, on noMMU targets, the host OS can be 'uclinux' instead of just
plain 'linux', which make sg3_util confused and thus it forgets to
include some source files in the build, resulting in link failures.
Fix that by ammending the regexp that recognises Linux. We now need to
run autoreconf.
Fixes:
http://autobuild.buildroot.org/results/79a/79a572f67bafa6a007256a5f22529c2daabd5733/http://autobuild.buildroot.org/results/4f7/4f7ea4bb4c4c12ceb9cd08bfb6720386f5bdea0b/
[Peter: fix typos as pointed out by Arnout]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
libintl may use thread-related functions, so -lpthread must come after
-lintl.
Fixes:
http://autobuild.buildroot.org/results/e9d/e9d89fd5b24a76a3c5ab9dc8f01f4643f7c13396/
[Peter: use -lpthread in comment as suggested by Arnout]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Add security related bugfix to fix CVE-2016-4429 from
GNU libc project.
Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
To avoid Xen build issues ensure that host Python 2 is used.
Signed-off-by: Alistair Francis <alistair.francis@xilinx.com>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
It uses spawn() and thus fork(), so it needs an MMU.
Fixes a build issue reported on IRC for a cortex-m4 build:
http://pastebin.com/dGCsy0sr
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
It uses GRND_NONBLOCK, introduced in 3.17.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
As described at:
4520524ba0
this commit continues a series of updates of ARC tools.
This time we're updating tools to arc-2016.09-eng008.
Main updates were made for gcc. It was switched to GCC 6 and to
OSABI v4.
Besides this patch fixes buildroot ARC failures connected to
"crtbeginT.o" object file missing. This issue lead to two main errors:
1) "crtbeginT.o: No such file or directory", e. g. bootutils-1.0.0.
No comments are required here I hope.
2) Errors like "compiler cannot create executables", e.g.:
a) host-gcc-final-arc-2016.09-eng007 static build,
b) aespipe-2.4c.
That was caused because the test to determine if compiler is able to
create executables was failing due to missing "crtbeginT.o" file.
We still keep GDB as it is of arc-2016.03 release because there're some
issues we'd like to resolve before releasing it to wider audience.
So again note this is next engineering builds of arc-2016.09 series
and it might have all kinds of breakages, please don't use it for
production builds.
Signed-off-by: Vlad Zakharov <vzakhar@synopsys.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
BR2_PACKAGE_UBOOT_TOOLS_FIT_SUPPORT selects BR2_PACKAGE_DTC, but forgets
to depends on !BR2_STATIC_LIBS like BR2_PACKAGE_DTC does.
Fixes the following kconfig warning:
arning: (BR2_PACKAGE_UBOOT_TOOLS_FIT_SUPPORT && BR2_PACKAGE_KVMTOOL && BR2_PACKAGE_UBOOT_TOOLS_MKIMAGE_FIT_SIGNATURE_SUPPORT) selects BR2_PACKAGE_DTC which has unmet direct dependencies (!BR2_STATIC_LIBS)
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This reverts commit 0d8158fc61. Thanks to
a fix in libaio, blktrace now builds fine on PowerPC, so there's no
reason to disable it anymore.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Both the blktrace and gadgetfs-test packages were failing to build on
PowerPC due to the mysterious:
hidden symbol `_rest32gpr_30_x' in libgcc.a(e500crtresx32gpr.o) is referenced by DSO
Due to this error, we disabled blktrace on PowerPC in commit
0d8158fc61. However, gadgetfs-test
continued to fail with the same error. As Romain Naour pointed out, the
problem seems in fact to come from a common dependency of blktrace and
gadgetfs-test: libaio. As Romain investigated, the problem started
appearing after the last bump of libaio, from version 0.3.109 to
0.3.110.
A quick bisect through the libaio changes between 0.3.109 and 0.3.110
has revealed that the problematic change is one in the libaio build
system, which now obeys to the CFLAGS provided in the environment,
rather than overriding them. So the CFLAGS provided by Buildroot cause
this problem. It turns out that the problematic CFLAGS is -Os, which is
indeed known to cause issues on PowerPC in some corner cases.
Even though it would probably be a better long-term solution to switch
to -O2 by default, and mark -Os as not available on PowerPC, it is a too
radical change so close to 2016.08. So we simply adjust the libaio
package so that it uses -O2 instead of -Os.
Fixes:
http://autobuild.buildroot.net/results/5e6cc4c432ce6c964ac285026978ad14d9eae97c/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
In the ffmpeg code, libavutil/arm/asm.S makes some assumption about the
ARM architecture for which the code is built. Only ARMv4, ARMv5, ARMv6
and ARMv7-A is supported. Due to this, object files built out of
ARM-optimized assembly code have the wrong architecture information,
causing a failure at link time.
Adding ARMv7-M support would be possible, but it doesn't exist yet in
ffmpeg, and it's pretty unlikely that ffmpeg will ever be needed on an
ARMv7-M platform, so this commit takes the simple approach of disabling
ffmpeg for ARMv7-M.
Fixes:
http://autobuild.buildroot.net/results/ca4c67b093afd6f14349fcdc87b02e0480172e8c/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
After fixing build issue related to MS_DIRSYNC and MS_RELATIME, we get
following build error with blackfin external toolchain.
fstest_monitor.c: In function 'main':
fstest_monitor.c:237: warning: implicit declaration of function 'fork'
/tmp/ccCzZIuW.o: In function 'main':
/home/rahul.bedarkar/buildroot/output/build/mtd-1.5.2/tests/fs-tests/utils/fstest_monitor.c:237: undefined reference to '_fork'
collect2: ld returned 1 exit status
make[2]: *** [fstest_monitor] Error 1
Since it uses fork(), add dependency on MMU support for integck.
Signed-off-by: Rahul Bedarkar <rahul.bedarkar@imgtec.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
libical requires time zone information in order to make use of time
zone functions (otherwise, segfaults occur when applications call
the time zone functions). This patch:
1. forces the system to build time zone info
2. ensures libical does not build its own time zone info (this is
the default behaviour, but the flag is added in case libical's
default changes in the future).
Signed-off-by: Marc Khouri <marc@khouri.ca>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Bump the version to get two portability patches (as well as a small,
innocuously-looking minor feature).
Fixes:
http://autobuild.buildroot.net/results/682/68266cb5b26a62387dc99aef31fc9654c8fcd505/
[and quite a few others...]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Maxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
As reported by Manuel in bug #9161, the commit
19b503b0d2 that bumped the X.org server
from 1.18.3 to 1.18.4 forgot to also rename the patch directory, making
the existing patch orphan, while it is still needed. This commit
therefore renames the 1.18.3 directory to 1.18.4.
Moreover, this patch is also needed for the 1.17.x X.org server series
which has been added recently, so this commit also adds the patch to a
new 1.17.4 directory.
Fixes bug #9161.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
As discussed with Waldemar, the C++ support for Blackfin is currently
broken, and we don't have a fix in sight for the 2016.08
release. Therefore, this commit disables C++ support entirely on the
Blackfin architecture in the internal toolchain backend.
This will avoid a significant number of Blackfin build failures, that
occur when building C++ packages.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
When readline is static library, we need to link against ncurses,
it is because, dependent library's symbols are not resolved when
static library is built. Those symbols are resolved when binary tries
to link with static library.
We can't pass linker flags for ncurses by setting LIBS environment
variable via <PKG>_CONF_ENV because it looks like build system is not
taking that into account and even though it would have been, order of
linking is important.
We can't also pass linker flags for ncurses via --with-readline-libs
conf options because it causes lftp_LIB_READLINE macro to take readline
headers from host machine if available. To use --with-readline-libs
we need to set --with-readline=yes and --with-readline-inc to include
dir. But when --with-readline=yes, readline_prefix is computed based
on if headers can be found in /usr/local or /usr. If readline is
installed on host machine, then configure fails since we are using
headers for host machine. If headers are not found in /usr/local or /usr
then only path specified --with-readline-inc is taken into account.
So specifying linker flags for ncurses via --with-readline-libs will
not work in all cases.
So it looks like, updating linker flags directly is only option to fix
static link issue against readline.
Fixes:
http://autobuild.buildroot.net/results/ae9/ae93bd1b4f6e1ed8f386830d4a51f00d9466f267/
Signed-off-by: Rahul Bedarkar <rahul.bedarkar@imgtec.com>
[Thomas: update autoreconf comment in .mk file.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
When installation prefix is specified in "--with-readline" config
option, lftp's build systems figures out include dir and linker flags.
So don't specify --with-readline-{inc,libs} options which are anyways
get ignored by build system.
Signed-off-by: Rahul Bedarkar <rahul.bedarkar@imgtec.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Fixes:
CVE-2016-5384 - possible double free due to insufficiently validated
cache files.
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Currently, we limit installing zoneinfo to non-musl toolchain, by lack
of knowledge on how it would work on musl.
Turns out that musl uses the same zoneinfo format as glibc does.
Make it possible to install the TZ info whatever the C library; for
musl, use tzdata as for glibc.
Thanks Rich! ;-)
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Rich Felker <dalias@aerifal.cx>
Cc: Marc Khouri <marc@khouri.ca>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This commit fixes several build issues of OpenBLAS on ARM:
- The first one occured on ARMv5 platforms, when the ARMV5 OpenBLAS
architecture is used. In this case, OpenBLAS build system forces
-march=armv5, which may not be correct for certain toolchains. As an
example, the Sourcery CodeBench toolchain has an ARMv4 and an ARMv5
sysroot. The ARMv5 sysroot is actually an armv5te sysroot, so when
OpenBLAS forces -march=armv5, gcc thinks it should use the ARMv4
sysroot, causing build failures.
To address this, a patch to completely remove the -march ARM CFLAGS
is added to OpenBLAS.
Fixes:
http://autobuild.buildroot.net/results/991497b12b70f948169e5ad99eebd0fe7f6209a2/
- The second one occured on ARMv7 platforms, when the ARMV7 OpenBLAS
architecture is used. The OpenBLAS code expects an EABIhf build, so a
dependency is added for EABIhf for both ARMv6 and ARMv7.
Fixes:
http://autobuild.buildroot.net/results/0ba0bee48a83367fcefab827e8eaa72f0c8fe90b/
- Once the previous ARMv7 problem has been fixed, it turns out that the
ARMv7 specific code in OpenBLAS contains VFPv3 specific
code. Therefore, the user *must* have choosen either VFPv3 or VFPv4,
or the code will not build. VFPv3-D16/VFPv4-D16 are not sufficient,
as more than 16 registers are used by the OpenBLAS code.
To address this, the ARMV7 platform of OpenBLAS is restricted to the
proper VFPv3/VFPv4 selection, and the ARMV6 platform is restricted to
the proper VFPv2 selection.
This problem was not visible in the autobuilders, as it was hidden by
the previous one.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This commit adds a new patch for mpv that fixes the stdatomic detection
logic so that it realizes linking against libatomic is needed on some
architectures. This fixes the build of mpv on SPARC. The patch has been
submitted upstream.
Fixes:
http://autobuild.buildroot.net/results/f607eed9a48c4ca00db43f9d2652e615b4bd7c93/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
While musl and glibc provide the ucontext structure definition on all
architectures they support, it is not the case of uClibc. We have been
excluding more and more architectures over time, but this isn't really
nice as we continuously need to add more. Latest example: the fact that
ucontext on ARM is only available on ARM platforms that support the ARM
instruction set (i.e not on the Thumb-2 only Cortex-M platforms).
Therefore, we introduce a BR2_PACKAGE_LIBSIGSEGV_ARCH_SUPPORTS variable,
which:
- is set to 'y' for musl and glibc
- is set to 'y' for uClibc for the only architectures that are known to
provide ucontext (those are the ones that select ARCH_HAS_UCONTEXT in
uClibc.
Note that we remove the comment, as it would become a way too
complicated comment, and we usually don't add comment for such
complicated situations (see libunwind for a similar example).
Fixes:
http://autobuild.buildroot.net/results/b1e40b3ec64cd98b535d83e89e5780c46680e095/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The dante configure script has some logic to remove the -g flag from
CFLAGS, but this logic is flawed and also removed -g in
-mfloat-gprs=double, turning it into the invalid -mfloatprs=double,
causing a build failure.
This commit adds a patch that gets rid of this -g removal logic, since
it is in fact unnecessary.
Fixes:
http://autobuild.buildroot.net/results/313370bf05efe7fd87c281a97ecb6e06531a87ed/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The procps-ng package currently fails to build on BR2_STATIC_LIBS=y
configurations. Indeed, by default NUMA support is enabled, and it
requires dlopen(). As hinted by the configure script, passing
--disable-numa allows to disable NUMA support.
However, once this is done, another issue pops up: dlopen() is also used
by the SELinux support. But even when SELinux support is disabled, the
procps-ng code incorrectly includes <dlfcn.h>. This is addressed by the
addition of a patch.
Fixes:
http://autobuild.buildroot.net/results/b385bf435085728aece6323a5006ba9fa6631744/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
webrtc-audio-processing assumes that execinfo.h is available when
__UCLIBC__ is not defined, which is an incorrect assumption, and
specifically not valid for musl.
This commit adds a patch that moves to autoconf based checks for
cxxabi.h and execinfo.h, which solves the problem.
Fixes:
http://autobuild.buildroot.net/results/480f529a53b321344e499f958be8e5d1d0f58646/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The webrtc-audio-processing configure script assumes that if the
host_cpu part of the tuple is "arm", then ARM instructions are
available. This is obviously incorrect for ARM Cortex-M platforms, which
only support the Thumb-2 instruction set.
In order to address this, we add a patch,
0001-configure.ac-fix-architecture-detection.patch, which changes how
webrtc-audio-processing detects the architecture: instead of relying on
the host_cpu part of the tuple, it relies on the built-in definitions of
the compiler.
Not only it fixes the Cortex-M detection, but it also enables ARMv7
optimizations on ARMv7-A: until now they were only enabled when the
host_cpu part of the tuple was armv7, which is never the case in
Buildroot.
However, once this issue is fixed for Cortex-M, the build nonetheless
fails later due to the usage of NPTL-only functions. So we change the
thread dependency to a NPTL dependency.
Fixes:
http://autobuild.buildroot.net/results/4933082cd7cc5781404c77ccef5c2b9333c5f714/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Otherwise conflicts with the naming convention from previous section.
Signed-off-by: Gary Bisson <gary.bisson@boundarydevices.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The size-stats script fails when the usb_modeswitch_data is enabled,
because this package installs files that contain commas in their
name. However, the size-stats script also uses comma as a separator for
its CSV files, causing a "ValueError: too many values to unpack" in:
pkg, fpath = l.split(",")
Fix this by splitting only the two fields that need to be split.
The bug was reported by Matthias <porto.rio@gmx.net>, who also suggested
a fix.
Fixes bug #9136.
Reported-by: Matthias <porto.rio@gmx.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
