system/timezone: check localtime exists

When installing a localtime, check it is a valid timezone.

[Peter: extend error message to make it clear to the user what to change]
Reported-by: Sagaert Johan <sagaert.johan@skynet.be>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

ruby: disable openssl for host variant

As reported by Fatih Aşıcı on the mailing list it sometimes fails with
the host openssl depending on versions when development files are
installed.
And as suggested by Thomas Petazzoni it's really not required for
anything so just disable it.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

dosfstools: fix static linking with libiconv

Fixes http://autobuild.buildroot.net/results/246/246c067f1e9def748498f0c6fa6988c036e1d109/

Pass the libraries to be linked in LDLIBS instead of LDFLAGS so the end up
after the object files on the linker command line.

While we are at it, use TARGET_CONFIGURE_OPTS instead of explictly passing
CC/CFLAGS/LDFLAGS.

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

flite: fix static linking against alsa

Fixes:
http://autobuild.buildroot.net/results/3ff/3ff2217a4fccbddbdb2bc8ade68c88fc52848d36/

Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libv4l: needs mmu

Fixes:
http://autobuild.buildroot.net/results/989/989d32f3532283817e9439aa78cceff5983dc421/

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

Update for 2014.05-rc2

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

wpa_supplicant: fix internal TLS implementation security issues

Add upstream patches fixing internal TLS validation of X.509 certificates. See
http://lists.shmoo.com/pipermail/hostap/2014-May/030273.html for details.

Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

infra: pkg-utils.mk replace "echo -e" with printf

printf is POSIX-compliant, echo -e is not.

Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

acpid: replace "echo -e" with printf

printf is POSIX-compliant, echo -e is not.

Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

toolchain-external: fix Linaro 2014.02 symlink

In commit cd32da8f7931af61737df293d1f8bcedaab8648e
("toolchain-external: add Linaro ARM big endian toolchain") a mistake
was made, probably due to a rebase conflict that was incorrectly solved:
the call to the post install staging hook that creates the necessary
symbolic links for a root filesystem based on Linaro 2014.02 to work
was removed.

This commit reinstates this call, which should fix the problem
observed by Maxime Hadjinlian while using Linaro 2014.02.

Reported-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Tested-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

nodejs: fix build when python3 is selected in Buildroot

When Python 3 is selected in Buildroot, the host/usr/bin/python
symlink of the host Python interpreter points to python3. Packages
that need to use the host Python 2 interpreter have to use python2.

In commit 40218a16526946f35dab0d05f379f71a01f29530 ("nodejs: force
python interpreter"), Samuel made some changes to the nodejs package
to use python2. One part of the changes is to sed a .gyp file to
replace the string 'python' by the path to python2. However, this
operation is done *after* calling the configure script, so it has in
fact no effect. Putting this sed before calling the configure script
fixes the problem.

However, there is a better solution: the nodejs build system has a
mechanism of variables, and it already defines a python variable
according to the environment variable PYTHON being passed. So this
patch instead adds a new patch to nodejs to use this python variable.

Fixes:

  http://autobuild.buildroot.org/results/aff/affd7300895ec400de50a33d51b4e94e15d63341/

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

packages: make providers declare what they provide

Fixes:
    http://autobuild.buildroot.org/results/285/2851069d6964aa46d26b4aabe7d84e8c0c6c72ce
    http://autobuild.buildroot.net/results/9b7/9b7870354d70e27e42d3d9c1f131ab54706bf20e
    [...]

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: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

docs/manual: document the new FOO_PROVIDES variable

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: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

infra/pkg-virtual: validate only one provider provides an implementation

Currently, it is possible that more than one provider of a virtual package
is selected in the menuconfig.

This leads to autobuild failures, and we do not protect the user from
making a mistake in the configuration. The failure is then hard to
troubleshoot in any case.

We can't use kconfig constructs to prevent this, since kconfig does not
tell how many options did a select on another option.

This change introduces a new variable a provider *must* define to include
all the virtual packages it is an implementation of. Then, when evaluating
the package's rules, we check that the provider is indeed the declared one
for each virtual package it claims to be an implementation of.

This works by taking advantage that when more than one provider is
selected, only one of them will 'win' in setting the _PROVIDES_FOO
option. Thus any provider just has to check it is indeed the declared
provider. If not, it means that one or more other provider is selected.

This gives the opportunity to the user to change its configuration, and
we can match the error message in the autobuilders to skip those failures
(we can skip them instead of reporting them, since they are obviously
configuration errors that should not happen in the first place.)

[Note: kudos to Arnout for suggesting this actual implementation. :-)]

Fixes:
    http://autobuild.buildroot.org/results/285/2851069d6964aa46d26b4aabe7d84e8c0c6c72ce
    http://autobuild.buildroot.net/results/9b7/9b7870354d70e27e42d3d9c1f131ab54706bf20e
    [...]

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: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

build/sstrip: deprecate for 2014.05

Quoting Thomas on IRC:
  > sstrip is a piece of crap that we should exterminate
  > from the surface of the planet

sstrip is now unmaintained, has a long history of generating butchered
binaries in some cases, and does not gain that much space in the end.

We'd like to obliviate sstrip from the strip options, anihilate the
sstrip package, scorch any reference in fire, and cleanse our tree
free of any remnants...

But first, let's just deprecate it for now.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Peter Korsgaard <jacmet@uclibc.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

uboot: check various configuration parameters and bail out if empty

Following the issues reported by Jerry <g4@novadsp.com>, it appears
that there are a number of U-Boot configuration variables we are not
checking properly, leaving users with weird behavior at build time.

This commit extends the existing U-Boot checks to verify that when a
custom version, custom tarball, or custom Git or Mercurial
repositories are selected, the appropriate fields are not empty in the
configuration.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

ipsec-tools: needs MMU support

ipsec-tools currently fails to build for Blackfin due to GLOB_TILDE
not being supported in Blackfin external toolchains. However, after
fixing this problem (which will be fixed in later versions of the
Blackfin external toolchains) allows to discover that ipsec-tools uses
fork() anyway.

Therefore, this patch simply marks ipsec-tools as not available on
non-MMU systems.

Consequently, it fixes:

  http://autobuild.buildroot.org/results/fc4/fc4ef4101a20e64eb52da56557d6996dd73b3d86/

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

matchbox-keyboard: fix dependencies

Correct a number of issues with matchbox-keyboard:
 - Missing dependencies on host-pkgconf and expat
 - Missing kconfig select on xlib_libXft
 - pango/cairo confusion, ensure cairo is built before if enabled
 - _CONF_OPT variable typo

Fixes:
http://autobuild.buildroot.net/results/473/47380d57e507ec395a1132dfd7b32d35459c1c7a//
http://autobuild.buildroot.net/results/d9a/d9aa2b6ddce920ea45fbd7093addfc0ee6b94467//
http://autobuild.buildroot.net/results/aad/aad393a81812df28cbc296f556de2acb4e6d114b//
http://autobuild.buildroot.net/results/124/124b28336ed61822fa6d5d2591b4e70d83b8a213//
and 17 other build errors, dating back to 2013-07-03

Reported-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

pciutils: disallow on Blackfin

Even though some Blackfin processors have PCI interfaces, the PCI
support is not supported by the Linux kernel: it has been marked as
broken since 2008. This means there's probably little interest for PCI
on Blackfin, and consequently little interest about getting pciutils
to build.

Fixes http://autobuild.buildroot.net/results/a27/a27ce61967ef8f780887f7681c8b0edd91656be6/

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

samba: disable on AVR32

samba produces particularly large binaries, and the relocations needed
for it do not fit in the possible relocation mechanisms available on
AVR32. Since samba is very unlikely to be used on AVR32, let's just
disable it, as we've already done for samba4.

Fixes:

  http://autobuild.buildroot.net/results/0142b1d5b8c3f20f1bff76ece646d901407f7490/

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/webkit: Fix OpenGL-related dependencies

fixes
http://autobuild.buildroot.net/results/27a/27a674c6a19a729fae5cc33de0360b2a061930c1/

This patch indeed fixes several problems based on the autobuilder-defconfig used.

1) Add missing dependencies when opengl is detected by webkit to satisfy autobuilder:

checking which GPU acceleration backend to use... opengl
checking for XCOMPOSITE... yes
checking for XDAMAGE... no
configure: error: Package requirements (xdamage) were not met:

2) ... but opengl should not have been detected with the used defconfig:
$ grep MESA .config
BR2_PACKAGE_MESA3D=y

which will lead to a link error although webkit configure found all dependencies:

/home/br2/output/host/opt/ext-toolchain/bin/../lib/gcc/mips-linux-gnu/4.7.3/../../../../mips-linux-gnu/bin/ld: cannot find -lGL

3) same is valid for glx, having GL/glx.h means nothing

4) add some configure options for gles/egl as well, just to make sure ;)

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gdb: switch to 7.x for Blackfin

For some reason, we were keeping gdb version 6.6a specially for
Blackfin. However, it turns out that support for Blackfin was merged
in gdb 7.4 (cross-gdb and gdbserver, not native gdb on the
target). Therefore, we can simply remove the support for version 6.6a
and use 7.5 as the default.

Both 7.4 and 7.5 were built tested, including cross-gdb and gdbserver.

The original reason to switch to 7.x is that 6.6a doesn't build for
Blackfin FLAT, and while it builds for Blackfin FDPIC, it only builds
libiberty.a and does not actually build a cross debugger.

Fixes:

  http://autobuild.buildroot.org/results/b47/b47f85553336b7f63ee4ecdf8598374ce4a225a3/

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gdb: fix build of host-gdb 7.4

The build of host-gdb 7.4 fails due to some texinfo issue. To avoid
that, use the same trick as is used for the target variant of gdb:
tell gdb that makeinfo is missing.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

fio: disable on nios2 because of missing fallocate64()

The fallocate() function is replaced by fallocate64() while compiling
with LARGE_FILE flags. Current nios2 toolchain is missing the fallocate64()
function.

Fixes:
    http://autobuild.buildroot.net/results/9cd/9cd30031a40f768f6090cfba44c880fb2406672b
    http://autobuild.buildroot.net/results/7da/7da32ca8178aa6bada06dfaea38e9e693fba0db2
    http://autobuild.buildroot.net/results/0ef/0ef7358289bff2a5f9efda27115eddb379c1a0b9
    [...]

[Peter: reference BR2_LARGEFILE, add autobuilder refs as suggested by Yann]
Signed-off-by: Frank Bergmann <frank@frajasalo.de>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

linux-headers: bump 3.{2,4,12}.x series

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

samba4: disable libbsd support

There's a symbol conflict between regular (POSIX) link(2) and the BSD
variant for builtin heimdal when libbsd is around and used. Fixes:
http://autobuild.buildroot.net/results/657/65726ceccbc2d8fa24e178ea66cd44703768bc95/

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

Makefile: move fs/common.mk above external.mk

The file external.mk was included before fs/common.mk, so it was impossible
to add rootfs targets using the BR2_EXTERNAL mechanism.
This change moves the inclusion of fs/common.mk before external.mk to allow
this.

Signed-off-by: Philippe Reynes <tremyfr@yahoo.fr>
[ThomasDS: rebased, update commit message]
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

news.html: add 2014.05-rc1 announcement mail link

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/imagemagick: remove useless makefile target.

Makefile's target "install-magickincarchHEADERS" and "magick-install-data-local"
install both the same file (magick-baseconfig.h) in the same time...

The problem can be reproduced with:
mkdir /tmp/bar
/usr/bin/install -c -m 644 foo /tmp/bar &
/usr/bin/install -c -m 644 foo /tmp/bar/foo
/usr/bin/install: cannot create regular file '/tmp/bar/foo' : File exists

So, remove one of them.

Fixes:
http://autobuild.buildroot.net/results/d34/d34077ce582866c50bbd90de10bbe593e39463f1/build-end.log

Signed-off-by: Romain Naour <romain.naour@openwide.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

duma: not available on non-MMU platforms

By design, building duma on Blackfin and other non-MMU platforms is
quite useless: it uses page mappings to detect buffer overrun and
use-after-free programming errors. Therefore, this commit simply marks
duma as not available on non-MMU platforms.

Fixes:

  http://autobuild.buildroot.org/results/951/9519d7abb45c6a789874ec6bb84418c31a98916b/

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

fbgrab: fix static linking

Fixes:

  http://autobuild.buildroot.org/results/9e0/9e079632a6c0556c96ec8ed7564286aa84044bf7/
  http://autobuild.buildroot.org/results/401/401105b5e120176e78c2e5fcabe54ab0bde1bf40/
  and more

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

mplayer: not supported on ARC

Fixes:

  http://autobuild.buildroot.org/results/af1/af1ca869dbc7a71aa4cc89373d605ca8f790fcc9/

[Peter: also adjust largefile comment dependency]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

cairo: fix build issue with gcc 4.9

Fixes:

  http://autobuild.buildroot.org/results/277/277038de62d6262576a63b213374a33357a8333b/

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

Drop 'foo' debug file from previous commit

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

graph-depends: rename the variable to pass the max depth

There will soon be new options to the graph-depends script, which we
can only sanely pass via environment variables.

Currently, we use such an environment variable to pass the maximum depth
of the dependency graph; the name of that variable is explicit that it
contains just the depth.

However, there has been so far no release of Buildroot which would make
use of that variable, so no user should have come to rely on it.

Rename that variable so it is less specific, and more generic, so it can
be used to pass more options to graph-depends.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/duma: do not let Makefile guess OS

We need to tell the duma Makefile what OS we're gonna run on, or it
uses default values.

It turns out the default values are the same as for OS=linux (and so
this change should be a no-op), but it's just better to force it, and
it avoids a warning during the build.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

xlib_libXpm: fix build on noMMU architectures

Add a patch to libXpm to fix build on noMMU architectures.

Fixes:

  http://autobuild.buildroot.org/results/c57/c571b0a9d9f36a445ea36a4bd609154f9a9bddd5/

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

pulseaudio: uses fork(), not available on noMMU platforms

Even though pulseaudio has some HAVE_FORK compile-time conditionals to
avoid using fork(), those parts of pulseaudio are not using it. Since
using pulseaudio on !MMU platforms is fairly unlikely, we simply
disable it when no MMU is available, and propagate this dependency to
the appropriate locations.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

fluxbox: uses fork(), not available on noMMU platforms

fluxbox uses the fork() system call, not available on noMMU
platforms. It's also fairly unlikely to have to use a full-blown
X.org window manager on such platforms.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

php-imagick: propagate BR2_USE_MMU dependency

php-imagick selects imagemagick, and imagemagick 'depends on
BR2_USE_MMU', but this dependency was not propagated to
php-imagick. This commit fixes that.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

zyre: propagate missing BR2_USE_MMU dependency

zyre selects filemq, and filemq 'depends on BR2_USE_MMU', but this
dependency was not propagated to zyre. This commit fixes that.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/x11r7/xserver_xorg-server: Udev kms support depends on libdrm

Fixes
http://autobuild.buildroot.net/results/18c/18c554d148fc00f85690c1638976dd6d0d7ff219/

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

mysql: Add optional OpenSSL dependency

Make sure OpenSSL is built before mysql.

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

mysql: Add optional zlib dependency

Make sure zlib is built before mysql.

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gpm: fix build on Microblaze due to ELF check

Fixes:

  http://autobuild.buildroot.org/results/19d/19d93324caac8b4aad7abfe7b6c3d85612636368/

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

sane-backends: uses fork(), not available on noMMU systems

sane-backends currently doesn't build with the ADI external
toolchains, due to an internal compiler failure. However, using the
internal toolchain backend for Blackfin goes past this internal
compiler failure, but raises the problem that sane-backends uses
fork(). Sane-backends has a --enable-pthreads option that allows to
use threads instead of forking in some parts of sane, but there are
still some other areas of sane-backends that continue to use fork().

Since the probability of using sane-backends on a noMMU system is
fairly small, we'll just disable this package for now on such
architectures, until someone cares enough to investigate deeper.

Fixes:

  http://autobuild.buildroot.org/results/136/136100c3a2d843b7915fdd83d23d8b301fee577e/

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

tstools: uses fork() therefore needs MMU

Fixes:

  http://autobuild.buildroot.org/results/4f6/4f611e085a3e21319a98f5383adf2caac77a0bd6/

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

elfutils: disable on Blackfin, problem of assembly symbol prefix

Fixes:

  http://autobuild.buildroot.org/results/922/92291286e86f9ea943ae7aa58cbd884c8db0c5d2/

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

elfutils: propagate architecture dependency to comment

elfutils depends on !BR2_avr32, but this dependency was not properly
propagated to the comment about toolchain options. This commit fixes
that.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

jack2: not available on ARC

ARC lacks some compiler intrinsics to build jack2.

Fixes:

  http://autobuild.buildroot.org/results/c21/c218be876d3abf7a8e212b6d526595bfd835e1c5/

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

jack2: requires MMU support

jack2 uses fork(), so it's not available on non-MMU architectures.

Fixes:

  http://autobuild.buildroot.org/results/0d4/0d441bf1831b2a657b466e44c29423c45c07decd/

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libdvd{nav, read}: add patches to fix build on NIOS II

Fixes:

  http://autobuild.buildroot.org/results/628/62839c4b2e687b940b97ad61bceb41cd5eb6e17a/build-end.log

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/avahi: does not build with static-only

... and propagate to reverse dependencies.

Fixes (which I reproduced here):
    http://autobuild.buildroot.net/results/ea2/ea2c8f20c4032bf82876805ea787a55834091cec

Should fix (which I did not even try to reproduce):
    http://autobuild.buildroot.net/results/4ef/4ef5bfc17f50bddb00f7ef3d62fc1fe963b7da82
    http://autobuild.buildroot.net/results/76e/76e4854812b0ef3862c080eead4012b35ef27f0f
    http://autobuild.buildroot.net/results/5bf/5bfb3fd9fe61bef5f998dc5999ce2060c334662c
    http://autobuild.buildroot.net/results/dc6/dc607c35471ef6dda04e06ad05ae7794615af578

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>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

udev: replace "echo -en" with printf

printf is POSIX-compliant, echo -e/n is not.

Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Acked-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

eudev: replace "echo -en" with printf

printf is POSIX-compliant, echo -e/n is not.

Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Acked-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

netsnmp: cleanup configure flags handling

No functional change, but cleaner.

Remove the double --disable-manuals and move the debug handling together
with the other optional configure flags.

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

netsnmp: Remove ldconfig calls

ldconfig breaks links created by gpu-viv-bin-mx6q.

fixes
http://autobuild.buildroot.net/results/ab5/ab57dd3729d529366808cbfc802f49360b2ec2b9/

[Peter: use NETSNMP_INSTALL_*_OPT, only install initscript for sysv init]
Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

qt5quick1: add missing dependency on network module

Signed-off-by: Fatih Aşıcı <fatih.asici@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

qt5quick1: removed OpenGL dependencies

Qt5Quick1 module doesn't need OpenGL libs and declarative module.

This module can be used to port qt4 qml apps to qt5. Patch has been tested
with linuxfb output plugin.

Signed-off-by: Wojciech Slenska <wojciech.slenska@gmail.com>
Reviewed-by: Fatih Aşıcı <fatih.asici@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libnss: Disable for MIPS64 n32 platforms

This package fails to compile for MIPS64 n32 platforms. We are in
conversations with upstream to fix this problem. Meanwhile we disable
this package until we fix it.

Fixes:
   http://autobuild.buildroot.net/results/0e3/0e3f1482d6f2f9bddc53d4e78b575120a2729e1d/

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

xserver_xorg-server: fix fontrootdir configuration

The --with-fontdir switch has been substituted by the --with-fontrootdir
switch a few versions ago. The use of the old configuration switch was
causing error messages like the following:

_FontTransOpen: Unable to Parse address ${prefix}/share/fonts/X11/*/
[dix] Could not init font path element ${prefix}/share/fonts/X11/*/, removing from list!

Signed-off-by: Filippo Sironi <filippo.sironi@gmail.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libunwind: depend on libatomic_ops when available

Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

luajit: fix a runtime error

this patch fixes :
    luajit: error while loading shared libraries: libluajit-5.1.so.2: cannot open shared object file: No such file or directory

Signed-off-by: Francois Perrad <francois.perrad@gadz.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gdb: add xtensa fixes for 7.5.1

Two fixes for xtensa are still applicable to gdb-7.5.1, which is now the
default; copy them from 7.4.1. This fixes native gdb build for xtensa.

Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gdb: replace xtensa patches with proper backports from binutils-gdb

The functionality is unchanged.

Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

jack2: force host-python dependency for waf build-system

jack2 uses Waf 1.6.11. The waf configure program combines a part of
python script with a compiled blob. According to Waf documentation,
waf currently supports python interpreter from version 2.3 to 3.1.

However, Buildroot provides python-3.4. So, let's uses host python2
interpreter for building jack2.

This patch also cleanup whitespace, replacing wrongly used spaces with
tabs.

Fixed:
  http://autobuild.buildroot.net/results/913/913bb1836076be9f201bf6641c7614708d6916bc/

Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

binutils: add patch to fix AArch64 glibc build issue

The autobuilders are not able to build glibc on AArch64 since we moved
to the internal toolchain backend as the default for AArch64. The
reason is that the cross-ld segfaults while linking of the glibc
utilities. Unfortunately, the issue has so far been impossible to
reproduce outside of the autobuilder CI loop itself: even on the same
machine, in the same chroot, the issue does not appear.

Today, Will Newton helped me analyze a bit further the problem. In the
logs of the build machine, we have found that the following segfault
notification matched the dates of the autobuilder failures:

/var/log/syslog:May 15 10:39:20 sd-56966 kernel: [767023.832598] ld[6846]: segfault at 1b2c002 ip 00000000004caf80 sp 00007fffaa308df8 error 4 in ld[400000+1b9000]
/var/log/syslog.2.gz:May 13 07:58:24 sd-56966 kernel: [585094.434264] ld[17761]: segfault at 154b002 ip 00000000004caf80 sp 00007fff0d400168 error 4 in ld[400000+1b9000]
/var/log/syslog.3.gz:May 12 20:24:08 sd-56966 kernel: [543558.227915] ld[14841]: segfault at 2706002 ip 00000000004caf80 sp 00007fffc08d41c8 error 4 in ld[400000+1b9000]
/var/log/syslog.4.gz:May 11 19:51:59 sd-56966 kernel: [455484.523033] ld[50162]: segfault at 1f00002 ip 00000000004caf80 sp 00007fff14a02328 error 4 in ld[400000+1b9000]
/var/log/syslog.4.gz:May 12 05:13:37 sd-56966 kernel: [489085.120089] ld[32067]: segfault at 2430002 ip 00000000004caf80 sp 00007fffc448db48 error 4 in ld[400000+1b9000]
/var/log/syslog.6.gz:May  9 15:20:38 sd-56966 kernel: [266948.197141] ld[13483]: segfault at ff0002 ip 00000000004caf80 sp 00007fff7e4fe948 error 4 in ld[400000+1b9000]
/var/log/syslog.7.gz:May  8 16:45:08 sd-56966 kernel: [185853.688607] ld[1757]: segfault at 20b1002 ip 00000000004caf80 sp 00007fffd07d5ae8 error 4 in ld[400000+1b9000]

See
http://autobuild.buildroot.org/?reason=glibc-2.18-svnr23787&arch=aarch64
for the autobuilder results.

So, the segfault always happens while accessing an address 2 bytes
after the beginning of a page: it looks like a buffer overflow, which
gets triggered because the object is placed towards the end of the
previous page, and crosses over the next, unallocated page.

Moreover, we looked at the 0x4caf80 address in an AArch64 cross-ld
built by Buildroot, and it turns out that it is part of the
iterative_hash() function. This nicely correlates with the patch that
Will Newton had pointed in
http://lists.busybox.net/pipermail/buildroot/2014-May/095881.html as a
fix for invalid memory accesses reported by Valgrind in the AArch64
cross-ld, specifically around the same area of 'ld'.

Therefore, even though we have not been able to reproduce the issue
outside of the autobuilders, and cannot confirm that the patch is
fixing the issue, I propose to integrate this binutils patch into
Buildroot. It is anyway an upstream binutils patch, which fixes a real
problem.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

postgresql: disable spinlock support for microblaze (el/be)

Fixes:
  http://autobuild.buildroot.org/results/686/6860988c9edb62d0d32cc6b0bd6af4f35ca0e343/

Signed-off-by: Peter Seiderer <ps.report@gmx.net>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

lsof: don't use host toolchain to determine target libc

The lsof Configure script builds a test program using the host toolchain to
test whether glibc is used. This test is broken in cross compilation
environment. Set LINUX_CLIB to avoid the test. This should give the correct
result even for non glibc toolchains, since all libc variants we support have
the netinet/tcp.h header.

Fixes:
http://autobuild.buildroot.net/results/a1f/a1f0572dbf968c21f70b35cefff7ef7a1d9a348a/

Cc: Benoît Thébaudeau <benoit.thebaudeau@advansee.com>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

pkg-stats: fix whitespaces

Remove some spaces before tabs and add the empty line at end of file.

Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
Cc: Jerzy Grzegorek <jerzy.grzegorek@trzebnica.net>
Cc: Luca Ceresoli <luca@lucaceresoli.net>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

pkg-stats: add luarocks perl python virtual infrastructures

Recognize the recently added infrastructures: luarocks, perl, python,
virtual.

Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
Cc: Jerzy Grzegorek <jerzy.grzegorek@trzebnica.net>
Cc: Luca Ceresoli <luca@lucaceresoli.net>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

mtr: drop unneeded 'call' from autotools-package invocation

Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
Cc: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

openpowerlink: add Intel I210 chip support

Signed-off-by: Romain Naour <romain.naour@openwide.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

openpowerlink: use git repository instead of zip archive

Since commit 6c5c08b854e4490697076ae3c5a9c587d8672c63,
openpowerlink package is rebuilt at every make call because
.stamp_downloaded is missing in the build directory.[D

The culprit is OPENPOWERLINK_EXTRACT_CMDS that remove and
replace the build directory.

unzip extract a directory "openPOWERLINK-V1.08.4" and
Buildroot expect a directory "openpowerlink-V1.08.4"

It is easier to use git repository instead of zip archive and
avoids rename the directory openPOWERLINK to openpowerlink.

Signed-off-by: Romain Naour <romain.naour@openwide.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

cairo: fix build of any2ppm if no fork().

The test program any2ppm can run as daemon. This feature can be disabled
at compile time, if the required headers are not present. However the
support for fork() is not checked.

Fixes http://autobuild.buildroot.net/results/4f2/4f271d55d7060d412aa336cfc2da3be3538d3594/.

Signed-off-by: Eric Le Bihan <eric.le.bihan.dev@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/rsyslog: needs MMU (fork)

fixes http://autobuild.buildroot.net/results/5c0/5c03bf3a1e4e922a18b40f854ff59d905a93a89f/

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/directfb: Disable dynload when BR2_PREFER_STATIC_LIB is enabled

fixes http://autobuild.buildroot.net/results/a74/a74b97053d7cad4fcce59f658a4b7b4340284b80/

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/rsyslog: Add missing optional dependency libcrypt

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

linux: bump to version 3.14.4

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

linux-headers: bump 3.{4.10.14}.x series

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libunwind: add patch to fix behavior for ARM < v6

Since libatomic_ops does not implement real atomic operations for
ARMv4 and ARMv5, libunwind must define AO_REQUIRE_CAS do indicate it
requires compare-and-swap operations, even if not available as real
atomic operations for the current architecture. In this case,
libatomic_ops will rely on emulated atomic operations, which also
require linking against libatomic_ops, which was until now not done by
libunwind.

This fixes the mysterious ltrace build issue:

  http://autobuild.buildroot.org/results/e1b/e1b330abfa2d80f3f30bc3359428ea429c690eb8/

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

openvpn: bump to version 2.3.4

Fixes TLS version interoperability issues.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

flann: disallow static library build

Even though flann build system uses CMake, it always tries to build a
shared library, even if BUILD_SHARED_LIBS=OFF is passed. It could
probably be fixed in flann's CMakeLists.txt, but that's not really
worth the effort if upstream doesn't support it.

Therefore, we simply disallow the flann package in purely static
builds.

Fixes:

  http://autobuild.buildroot.org/results/a4e/a4ec0e9f28cd12ce770718fb407bbb4dc93b528b/

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

clapack: disable on blackfin

The "_" symbol prefix added by the blackfin toolchain causes "undefined
symbol" build failures when the compiler is trying to link a binary
with libblas and/or liblapack.

This patch also disables clapack reverse dependencies: armadillo.

Fixes:
  http://autobuild.buildroot.net/results/a28df0ba10803e6e54c2f8160bbd8190cba4d690/

Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

btrfs-progs: disable for avr32

Toolchain issues prevent it from being built, also there's no AVR32 with
proper storage ports for such a filesystem. Fixes:
http://autobuild.buildroot.net/results/06501fd32c30549810c12f0ad50f510c4b74ee63/

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

Update for 2014.05-rc1

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

wpa_supplicant: detect readline support in wpa_cli

Signed-off-by: Oscar Hellström <oscar.hellstrom@vinnter.se>
Acked-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

crda: disable on NIOS II

Building crda on NIOS II triggers the infamous _gp linker issue. Since
nobody cared, let's disable this package on NIOS II for the moment.

Fixes:

  http://autobuild.buildroot.net/results/272/272f18410e4855ab6b3bfd8a44243c1ebecf6abe/

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

cppcms: needs NPTL support

cppcms uses some thread functions that are not available in certain
uClibc thread implementations. Even though a bit more restrictive than
necessary, adding a NPTL dependency is the easiest solution, and is
quite logical for a relatively large and complex package such as
cppcms.

Fixes:

  http://autobuild.buildroot.net/results/a26/a26574419aacbea4140dfca1d503bcab599edd71/

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

CHANGES: update with recent changes

And move the xbmc packages to the new packages list, as that's what they
are.

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

aiccu: not available on NIOS II

Building aiccu does not work on NIOS II due to the _gp problem, and
nobody is interested to fix it for now, so we simply disallow this
package. While we're at it, we propagate the AVR32 exclusion to the
comment.

Fixes:

  http://autobuild.buildroot.net/results/786/7864cd5789a9f9b9f0603d738f4e63822d8469dd/

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

toolchain-external: Sourcery CodeBench ARM 2013.05 affected by PR58595

Fixes:

  http://autobuild.buildroot.org/results/6b3/6b342c2d101348e4d39285f45481e2aafa3d8b3f/

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

mesa3d: Fix gbm related compile error

https://bugs.freedesktop.org/show_bug.cgi?id=78225#c1
"AFAICS both gbm backends require DRI"

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

mesa3d: use --enable-shared-glapi also for Gallium drivers

Needed since this upstream commit:
http://cgit.freedesktop.org/mesa/mesa/commit/configure.ac?h=10.2&id=0432aa064bf5d4d0ad8fc3c4d648b8feb238ddfa

Remove --disable-shared-glapi from the non-DRI-block, this
would break with enabled Gallium drivers.

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

binutils: backport R_XTENSA_DIFF* relocation signedness fix

This fixes the following build errors:

  qlatincodec.o:(.debug_line+0xcd7): dangerous relocation: overflow after relaxation
  qmetaobject.o:(.debug_frame+0xa89): dangerous relocation: overflow after relaxation

Fixes:
  http://autobuild.buildroot.net/results/f51/f51c87ae3c883cc36b06c0741999b52e48e4c943/
  http://autobuild.buildroot.net/results/bcb/bcbf25e4054911a7a192bdb5c92e0ebd50e9666c/
  http://autobuild.buildroot.net/results/8a8/8a8a0f8cca293ce1f383421ba8b1223b4266e3bc/
  http://autobuild.buildroot.net/results/a54/a54f3707781dc193292eda7d1723fce6487d6980/

Backported from: 1058c7532d0b012ac329219264ddad59049fb6e6
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

binutils: backport first section frag alignment fix for xtensa gas

This fixes the following build error often seen during linux kernel
linking step with relaxation enabled:

  dangerous relocation: call8: misaligned call target: (.text.unlikely+0x63)

Backported from: a35d5e823fdfe8a6e7e05ca8e3fb8bb5697335b1
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

manual: add information about INSTALL_IMAGES_CMDS

The Buildroot manual currently does not contain any mention of the
<pkg>_INSTALL_IMAGES_CMDS which could potentially be of interest when
developing a new 'generic-package'.

Signed-off-by: Ryan Barnett <rjbarnet@rockwellcollins.com>
CC: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/xbmc: add missing dependency on libtheora

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Bernd Kuhls <berndkuhls@hotmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

mesa3d: Add patch note about upstream bug

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

sysstat: fix passing CFLAGS to make

SYSSTAT_MAKE_OPT is a string used to pass options to make, '+=' operator
doesn't have any special meaning inside it, so CFLAGS+=... is passed to
shell, overwriting previous CFLAGS value.

Replace CFLAGS+="..." with CFLAGS="$(TARGET_CFLAGS) ...".

Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Reviewed-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>