arch: remove BR2_arm10t

The BR2_arm10t option is not correct as it references an ARM family,
while other options indicate a specific ARM core. The ARM cores in
ARM10 family are ARM1020E, ARM1022E and ARM1026EJ-S according to
Wikipedia. However, those are clearly very rare, and Wikipedia only
indicates two Conexant ADSL-related SoC as being part of this family
of ARM cores. Therefore, this commit removes this ARM family.

[Peter: remove nettle.mk reference as pointed out by Yann]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

arch: remove BR2_arm920 reference

The BR2_GCC_TARGET_CPU defines a value for the BR2_arm920 case, but
this option does not exist. Therefore, this commit removes one line of
dead code.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gnutls: bump to version 3.2.18

Also add hash, calculated on downloaded file after verifying signature.

[Peter: tweak hash comment]
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libpfm4: needs NPTL support

Fixes following build failure:
--->---
self_smpl_multi.c: In function 'my_thread':
self_smpl_multi.c:354:2: error: implicit declaration of function 'pthread_barrier_wait' [-Werror=implicit-function-declaration]
  pthread_barrier_wait(&barrier);
  ^
self_smpl_multi.c: In function 'main':
self_smpl_multi.c:450:2: error: implicit declaration of function 'pthread_barrier_init' [-Werror=implicit-function-declaration]
  pthread_barrier_init(&barrier, 0, max_thr+1);
  ^
cc1: all warnings being treated as errors
--->---

And since xtensa, avr32 and bfin don't support NPTL we may remove arch
dependencies as well.

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

strace: Apply an upstream patch to fix sa_restorer problems

Wrap sa_restorer member definitions in #ifdef SA_RESTORER to be
consistent with their use.
If an architecture does not provide sa_restorer members but still
defines SA_RESTORER macro, the latter has to be explicitly undefined.

This change fixes compilation failures like this one:
signal.c: In function 'decode_old_sigaction':
signal.c:631:21: error: 'struct old_sigaction' has no member named
'sa_restorer'
signal.c: In function 'decode_new_sigaction':
signal.c:1224:21: error: 'struct new_sigaction' has no member named
'sa_restorer'

* signal.c (struct old_sigaction, struct old_sigaction32,
struct new_sigaction, struct new_sigaction32):
Wrap sa_restorer member in #ifdef SA_RESTORER.
(decode_old_sigaction, decode_new_sigaction):
Wrap use of sa32.sa_restorer in #ifdef SA_RESTORER.

Upstream commit:
  https://sourceforge.net/p/strace/code/ci/c3a5c01051ba29a9f421f4d45d96bbc32df2b796/

Fixes:
  http://autobuild.buildroot.net/results/93e/93e54d8aecc5c178cc3465eb9fd5415461325285/

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

e2fsprogs: switch site and add hash

For some reason the kernel.org download hash doesn't match the
sourceforge hash so switch to kernel.org for the download.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

iproute2: add hash

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libnl: add hash

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

openssl: add hash

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

wireless-regdb: add hash

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

btrfs-progs: add hash

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

iw: add hash

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

util-linux: add hash

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

iptables: add hash

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

polarssl: add hash

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

busybox: add hash

And rename 0006-lzop-add-overflow-check.patch to the proper convention.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

dialog: fixup ncurses config script

Point the build to the proper ncurses config script by using
$(NCURSES_CONFIG_SCRIPTS) from the ncurses package directly.

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

alsa-utils: fixup ncurses config script

Point the build to the proper (which BTW was broken since it pointed to
the wrong directory) ncurses config by using $(NCURSES_CONFIG_SCRIPTS)
from the ncurses package directly.
Also select the proper variant (widec/non-widec) in configure.

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

httping: enable ncurses support

Enable automatic ncurses support now that we've got wide support.

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

nano: enable ncursesw support

Enable support for ncurses widechar by specifying the proper
ncursesw-config when it's enabled, otherwise keep the old trick in place
when it's not to avoid automatically picking up the host/distro one.

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

package/ncurses: fixup library symlinks

As stated on the list we need to copy static libraries when doing static
targets so add the logic for that.

Also exclude the wide option for blackfin flat since there seem to be
toolchain issues with that combination - since it's a new feature option
someone interested might look into it later.

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

package/ncurses: Allow building wide char support

Allow ncurses to be configured with wide char support; this causes the
libraries to be built with the 'w' suffix (eg libncursesw.so,
libmenuw.so, etc), so we need to create a few symlinks.

Signed-off-by: Jeremy Kerr <jk@ozlabs.org>
Tested-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

linux: bump default to version 3.16.3

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

linux-headers: bump 3.{16, 14, 10}.x series

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

sysstat: bump to version 11.0.1

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

package/nodejs: bump to version 0.10.32

Signed-off-by: Jörg Krause <jkrause@posteo.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

dbus: security bump to version 1.8.8

Fixes:
CVE-2014-3635 - Buffer access with incorrect length value
CVE-2014-3636 - Allocation of file descriptors or handles
without limits or throttling
CVE-2014-3637 - Missing release of file descriptor or handle after
effective lifetime
CVE-2014-3638 - Algorithmic complexity
CVE-2014-3639 - Allocation of file descriptors or handles without
limits or throttling

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

ARC: gcc - Fix SIZE_TYPE to be "unsigned int" instead of "long unsigned int"

This makes size_t to be "unsigned" ssize_t which makes happy compiler on data
type checks.

Fix is taken from current development branch of GCC for ARC and will be a
part of the next release of ARC tools, so at that point patch should be dropped.

https://github.com/foss-for-synopsys-dwc-arc-processors/gcc/commit/249f040299402647525c3f15b79d319fa7acddd3

Fixes http://autobuild.buildroot.net/results/405/405da9a945511329929b18740b983c51b8dcc43e

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Anton Kolesov <akolesov@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Reviewed-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

sqlcipher: fix static linking

It uses openssl which for buildroot mandates libz, but doesn't link
against it so it fails. Fixes:
http://autobuild.buildroot.net/results/514/5145617f7c3cece933c845da3c9836d80d062bb6/

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

gst1-plugins-bad: fix sndfile comment

The package is not called gst-plugins-bad, and the other comments in the
file are of the "foo plugin needs a toolchain w/.." form, so use that for
sndfile as well.

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

gst-plugins-bad-plugin-sndfile: Fix comment display

Signed-off-by: Jérôme Pouiller <jezz@sysmic.org>
Reviewed-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gdb: add support for Python in target gdb

This commit adds a new option BR2_PACKAGE_GDB_PYTHON to enable Python
support in the target gdb. Since we can assume that the user will be
aware that Python is needed to get Python support in gdb, we chose to
use a "depends on" dependency instead of a "select" dependency.

The other weird thing is the need for a wrapper shell script to
replace gdb's provided python-config.py script. See the shell script
comment itself for all the details.

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

gdb: add support for Python in host gdb

This commit adds an option BR2_PACKAGE_HOST_GDB_PYTHON that allows to
enable Python support in the cross gdb built by Buildroot.

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

gdb: reword prompt for host TUI option

There is no need to name the option "GDB TUI support", since this
option is already visible "below" GDB in menuconfig/xconfig. Naming it
"TUI support" is therefore sufficient.

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

gdb: enable tui support for target package

Add a configuration option to compile the gdb target package with the
--enable-tui switch.

This is done pretty much in the same way as in commit 2474fb0bf1a7 ("host-gdb:
enable terminal user interface support"), but for the gdb package on target.

This makes sense only when a full debugger is installed on target,
not for a gdbserver.

[Thomas: remove "default n" since this is the default, adjust the
prompt of the option, and rewrap the help text.]

Signed-off-by: Vincent Stehlé <vincent.stehle@laposte.net>
Cc: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gdb: needs host-ncurses on the host

Commit 2474fb0bf1a786f3a170631f3dbce8b38092f986 ("host-gdb: enable
terminal user interface support") has added TUI support to host gdb,
and therefore added a dependency on host-ncurses when TUI support is
enabled.

However, host-ncurses is not only needed for TUI support, it is needed
for gdb in all cases as well, so this commit adds a dependency of
host-gdb to host-ncurses.

Fixes:

  http://autobuild.buildroot.org/results/153/153dbdc42103074f7a0895e8871e2eee4eae3325/

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

gdb: remove version 7.6

Now that the default version has changed to 7.7, we can get rid of the
older gdb 7.6 version.

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

gdb: switch to 7.7 as the default version

Now that gdb 7.8 is out, it's time to move to 7.7 as the default
version instead of 7.6.

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

gdb: add version 7.8

This commit adds support for the 7.8 version of gdb. Note that the
tarball of this version is not available as a .tar.bz2, so we have to
add a special case and download the .tar.xz for this version.

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

systemd: bump to version 216

Bump systemd to version 216. This new version provides two new tools to
manage the journal (systemd-journal-upload and systemd-journal-remote)
which resulted in the addition of new users.

Also remove backported patches.

Signed-off-by: Eric Le Bihan <eric.le.bihan.dev@free.fr>
Tested-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

p11-kit: bump to version 0.20.6

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

toolchain-external: bump ARM, ARMeb and AArch64 Linaro toolchains to 14.08

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

toolchain-external: remove CodeSourcery ARM 2012.03, add 2014.05

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

toolchain-external: add CodeSourcery AArch64 toolchain

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

lttng-babeltrace: Fix comment display

Signed-off-by: Jérôme Pouiller <jezz@sysmic.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

memtester: fix rebuild

Do the edit of conf-* files only when 'cc' appears at the beginning of line,
i.e. the files has not been edited. Otherwise, the 'cc' part of the cross
toolchain gets expanded, leading to the following error on 'make
memtester-rebuild':

./compile: line 3: /home/baruch/git/buildroot/output/host/usr/bin/i686-pc-linux-gnu-g/home/baruch/git/buildroot/output/host/usr/bin/i686-pc-linux-gnu-gcc: No such file or directory
Makefile:82: recipe for target 'memtester.o' failed
make[1]: *** [memtester.o] Error 127

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

celt051: Fix file naming

In order to comply with naming policy, fix celt.mk filename.

Signed-off-by: Jérôme Pouiller <jezz@sysmic.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libtasn1: bump to version 4.2

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

radvd: bump to version 2.7

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

qt5: bump to 5.3.2

Only use ccache prefix in QMAKE_CC and QMAKE_CXX since the build system is
broken when QMAKE_AR contains a space character.

Remove the upstreamed uClibc patch.

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

ipset: bump to version 6.22

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

gst1-plugins-good: gstv4l2allocator: O_CLOEXEC needs _GNU_SOURCE

Fixes http://autobuild.buildroot.net/results/ee9/ee985d94da6a7115b6e62cf407e3c6a90eb3e4bf/

On some systems (E.G. uClibc and older Glibc versions), O_CLOEXEC is only
defined when _GNU_SOURCE is specified, so do so.

Patch submitted upstream: https://bugzilla.gnome.org/show_bug.cgi?id=736670

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

qwt: add optional opengl support

qwt should only build the opengl support code if Qt is configured with
opengl support, so add a suboption for it similar to svg/mathml with the
needed dependencies.

Fixes http://autobuild.buildroot.net/results/597/5977bd77342e28ae9783f9f9edfcf84eefd081b6/

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

hostname: fix conflicting installs of /bin/hostname

Surprisingly long-standing issue with conflicting /bin/hostname
installs.  Reported as early as November 2005 by Joseph Dupre.

All together at one point or another there are at least 4 possible
sources of /bin/hostname:

        busybox
        util-linux
        coreutils
        net-tools

Buildroot depends on the -F flag being available in the default
/etc/inittab.  Out of the 4 listed projects only net-tools and buildroot
for sure support the -F flag.  I'm a little unclear on util-linux as it
has been removed entirely (in favor of net-tools) for some time.

As of coreutils 6.9.90 (2007-12-01), coreutils does not install its
/bin/hostname by default.  The following commit reenabled its build:

d6e58cb coreutils: fixed missing hostname (Sep 2010)

This was done to fix a build error in coreutils regarding help2man.  A
later patch:

30c5105 coreutils: bump to version 8.21

disabled the help2man functionality entirely but left hostname being
installed.

On a very related note, net-tools now contains an obsolete check to add
util-linux as a dependency to force it to build first (so that net-tools
ends up with /bin/hostname).

This patch fixes both of these issues so that hostname always comes from
one of two places:

    busybox
    net-tools

Tested-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Co-authored-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Andrew Ruder <andrew.ruder@elecsyscorp.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

Revert "package/localedef: don't use config.site when cross-compiling"

CONFIG_SITE is now a part of pkg-autotools infrastructure.

This reverts commit 85448febb3c9f21e6e4082fc344b3ef46261c782.

Signed-off-by: Romain Naour <romain.naour@openwide.fr>
Acked-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/pkg-autotools: don't use config.site when cross-compiling

On fedora 20 64bits host, the file /usr/share/config.site contains
a fix for installing libraries into /lib/lib64 on 64bits systems
that redefine libdir in the generated Makefile

For safety and avoid the bug #7262 [1], disable loading this file
when running the configure script for the target and the host.

Note: configure scripts generated with autoconf < 2.65 will source
the /dev/null and print this line:
"configure: loading site script /dev/null"

[1]: https://bugs.busybox.net/show_bug.cgi?id=7262

Signed-off-by: Romain Naour <romain.naour@openwide.fr>
Acked-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libqrencode: needs host-pkgconf

The configure script uses PKG_CHECK_MODULES to check for libpng.

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

dialog: Override ncurses config script

Without the override, systemwide ncurses6-config can be found instead
of ncurses5-config from staging.

Signed-off-by: Maarten ter Huurne <maarten@treewalker.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

google-breakpad: bump to version r1373

This version adds support for installing important header files to
<staging>/usr/include/breakpad. It's no longer necessary to include the
whole breakpad source tree when building applications using libbreakpad.a.

Signed-off-by: Frank Hunleth <fhunleth@troodon-software.com>
Reviewed-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

uclibc: two-stage gcc simplifications

After switching to a two stage gcc solution, there is no longer a need
to do weird things in the uclibc build. We can greatly simplify
UCLIBC_CONFIGURE_CMDS to only do the configuration, and let the
existing UCLIBC_BUILD_CMDS do the build. Note that we have to build
the headers before starting the C library build, otherwise there is a
build failure (probably a uClibc bug).

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

glibc: two-stage gcc simplifications

After switching to a two stage gcc solution, there is no longer a need
to do weird things in the glibc build. We can greatly simplify
GLIBC_CONFIGURE_CMDS to only do the configuration, and let the
existing GLIBC_BUILD_CMDS do the build.

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

musl: two-stage gcc simplifications

After switching to a two stage gcc solution, there is no longer a need
to do weird things in the musl build, with certain things being done
twice (MUSL_CONFIGURE_CALL). Now the MUSL_CONFIGURE_CMDS variable only
does the configuration, and the MUSL_BUILD_CMDS only does the build,
as it should be.

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

gcc/gcc-intermediate: remove package

Now that we have switched to a two steps gcc build process that uses
only gcc-initial and gcc-final, we can get rid of the gcc-intermediate
package.

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

toolchain: switch to a two stage gcc build

Currently, the internal toolchain backend does a three stage gcc
build, with the following sequence of builds:

 - build gcc-initial
 - configure libc, install headers and start files
 - build gcc-intermediate
 - build libc
 - build gcc-final

However, it turns out that this is not necessary, and only a two stage
gcc build is needed. At some point, it was believed that a three stage
gcc build was needed for NPTL based toolchains with old gcc versions,
but even a gcc 4.4 build with a NPTL toolchain works fine.

So, this commit switches the internal toolchain backend to use a two
stage gcc build: just gcc-initial and gcc-final. It does so by:

 * Removing the custom dependency of all C libraries build step to
   host-gcc-intermediate. Now the C library packages simply have to
   depend on host-gcc-initial as a normal dependency (which they
   already do), and that's it.

 * Build and install both gcc *and* libgcc in
   host-gcc-initial. Previously, only gcc was built and installed in
   host-gcc-initial. libgcc was only done in host-gcc-intermediate,
   but now we need libgcc to build the C library.

 * Pass appropriate environment variables to get SSP (Stack Smashing
   Protection) to work properly:

    - Tell the compiler that the libc will provide the SSP support, by
      passing gcc_cv_libc_provides_ssp=yes. In Buildroot, we have
      chosen to use the SSP support from the C library instead of the
      SSP support from the compiler (this is not changed by this patch
      series, it was already the case).

    - Tell glibc to *not* build its own programs with SSP support. The
      issue is that if glibc detects that the compiler supports
      -fstack-protector, then glibc uses it to build a few things with
      SSP. However, at this point, the support is not complete (we
      only have host-gcc-initial, and the C library is not completely
      built). So, we pass libc_cv_ssp=no to tell the C library to not
      use SSP support itself. Note that this is not a big loss: only a
      few parts of the C library were built with -fstack-protector,
      not the entire library.

 * A special change is needed for ARC, because its libgcc depends on
   the C library, which breaks building libgcc in
   host-gcc-initial. This looks like a bug in the ARC compiler, as it
   does not obey the inhibit_libc variable which tells the compiler
   build process to *not* enable things that depend on the C
   library. So for now, in host-gcc-initial, we simply disable the
   build of libgmon.a for ARC. It's going to be built as part of
   host-gcc-final, so the final compiler will have gmon support.

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

libqrencode: link with pthread to avoid build failure

Fixes: http://autobuild.buildroot.net/results/1ca0719dd1fe0bbc528ec99b5bc2b42e985bfdca/
[Peter: only if toolchain has threads, pass --disable-thread-safety otherwise]
Signed-off-by: Danomi Manchego <danomimanchego123@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

util-linux: disable python for host variant as well

It's bindings, and there's no need/use for them. Fixes:
http://autobuild.buildroot.net/results/f3d/f3dfe6bd9031645fb068385af4054c0cc7522b2b/

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

dos2unix: needs host-gettext

For msgfmt, fixes:
http://autobuild.buildroot.net/results/86a/86a104b4bfffa7426de93d63278b998d6dcce154/

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

uclibc: fix duplicate vfork problems when static linking

Reported on uClibc mailinglist by Thomas Petazzoni
Should fix:
* SuperH
 http://autobuild.buildroot.org/results/a81/a810c2f27dee1978808461c05cbdcbf51a214e09/build-end.log
* i386
 http://autobuild.buildroot.org/results/f83/f83b65bfc6ea7c7406a02e92afda43e4c5db6e7c/build-end.log
* x86-64
 http://autobuild.buildroot.org/results/64f/64f418f54885e8b5093dc8949c2d1d1ff3c938ea/build-end.log
* powerpc
 http://autobuild.buildroot.org/results/d5b/d5b7ac84dde0e4e26cd7cc46d79e66c39ed5cd53/build-end.log

Tested with i386 static build of alsa-utils.
Patch will be send to uClibc soon, need to finish test-suite run.

Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Tested-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

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

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

libelf: Removes the package

The elfutils package provides a more recent version of the
libelf. Some packages such as ltrace need this more recent
version. Having two packages providing the same library leads to some
conflicts or dependency problems. For instance at the end we had only
one libelf.a when the 2 packages were selected.

Signed-off-by: Gregory CLEMENT <gregory.clement@free-electrons.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

kexec-lite: Depends on elfutils instead of libelf to get the libelf library

The elfutils package provides a more recent version of the libelf, so
let's use it. It will allow to remove the libelf package and to avoid
conflicts with two packages providing the same library.

[Peter: adjust toolchain options comment to match]
Signed-off-by: Gregory CLEMENT <gregory.clement@free-electrons.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

avrdude: Depend on elfutils instead of libelf to get the libelf library

The elfutils package provides a more recent version of the libelf, so
let's use it. It will allow to remove the libelf package and to avoid
conflicts with two packages providing the same library.

[Peter: adjust toolchain options comment to match]
Signed-off-by: Gregory CLEMENT <gregory.clement@free-electrons.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

ltrace: Removes the libelf dependency

The libelf is currently provided by 2 packages libelf and
elfutils. The first package provides an old version of the libelf
which is no more compatible with a recent version of ltrace. This
patch removes the dependency on the libelf package and only keep the
elfuils package which provides the accurate version of libelf for
ltrace.

It will also allow to remove the libelf package and to avoid conflicts
with two packages providing the same library.

Signed-off-by: Gregory CLEMENT <gregory.clement@free-electrons.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

procps-ng: Fix installation and take precedence over busybox

Currently procps-ng binaries are installed under $(TARGET_DIR)/usr/usr
instead of $(TARGET_DIR)/usr, which would be the expected. This is
caused because the default value of the --exec-prefix configure option
is /usr, and Buildroot always adds --prefix=/usr to the configure
option. We fix this by setting --exec-prefix to an empty value.

At the same time we make sure the procps-ng binaries have precedence
over the busybox ones.

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Reviewed-by: Markos.Chandras <Markos.Chandras@imgtec.com>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Tested-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

coreutils: get the uptime program to work

coreutils configure script needs to know whether /proc/uptime is
available, but this is not possible in a cross-compilation
context. This leads to an uptime program that fails to work on the
target, as it has been compiled without /proc/uptime reading support.

This commit fixes that by telling coreutils at configure time that
/proc/uptime will be available on the target (which seems to be a
reasonable assumption on Linux systems).

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Tested-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

dos2unix: new package

Signed-off-by: David Bachelart <david.bachelart@bbright.com>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Tested-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

sqlite: add option to disable fsync

For use-cases in which performance is more important than data integrity.

Signed-off-by: Floris Bos <bos@je-eigen-domein.nl>
Reviewed-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

openvmtools: new package

New package: openvmtools

[Peter: add patch to fix compilation against modern glib versions]
Signed-off-by: Karoly Kasza <kaszak@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

tn5250: fix openssl dir

If openssl devfiles are installed in the host the automatic lookup might
find that and cause issues.
So be explicit on the directory to avoid a search and patch the
directory logic since it's got a bug. Fixes:
http://autobuild.buildroot.net/results/5dd/5ddc8636d6555851f692c0c21f583df235340b19/

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

util-linux: bump to version 2.25.1

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Reviewed-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

openssh: Add patch to fix compilation with musl libc

The configure script finds the "howmany" macro, but some of the sources
using it do not include the required <sys/param.h> header.

Signed-off-by: Maarten ter Huurne <maarten@treewalker.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

strace: Disable for MIPS when using a uClibc toolchain

A change introduced in strace-4.9 makes it incompatible with MIPS
architecture when using a uClibc toolchain. This is the failure:

signal.c: In function 'printsiginfo':
signal.c:453:9: error: 'siginfo_t' has no member named 'si_timerid'
      sip->si_timerid, sip->si_overrun);
 ^
signal.c:453:26: error: 'siginfo_t' has no member named 'si_overrun'
      sip->si_timerid, sip->si_overrun);

uClibc needs to be patched in order to fix this problem. In the meantime
we disable strace for MIPS when using a uClibc toolchain.

Fixes:

  http://autobuild.buildroot.net/results/91c/91c0413e755c365332bc15d1d69ab3d6160ae78c/

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

toolchain/musl: disable for PowerPC SPE

It's not supported and the build breaks.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

ipmitool: new package

Adds ipmitool, a command-line interface to IPMI-enabled devices.

Signed-off-by: Floris Bos <bos@je-eigen-domein.nl>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libsvg-cairo: explicitly disable X support

Our libsvg-cairo package doesn't have any dependency on X, and the
source code doesn't seem to be using X. However, the configure script
checks for X headers and libraries. If they are found on the host,
/usr/lib and /usr/include get added to the library/header paths
causing weird build failures. This commit should fix this problem by
explicitly disabling the X support.

Should fix:

  http://autobuild.buildroot.org/results/d0f/d0fc233bb97c2ea18344746a5b7c63c8de65330a/
  http://autobuild.buildroot.org/results/0b3/0b3fb2ee9916aa265df5b6e4f04f6dfa25d1fc84/

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

erlang: needs thread support

Fixes:

  http://autobuild.buildroot.org/results/1f3/1f313aee75328c27e260c55dbf7caa5d3d6a07d3/
  http://autobuild.buildroot.org/results/2f2/2f213d8cf6f828a68a2f36bb95564b8db043e7a2/

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

heirloom-mailx: add license info

[Peter: drop unused IEEE part as suggested by Baruch]
Signed-off-by: Danomi Manchego <danomimanchego123@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

lua: fix license path for lua 5.2

The file containing the license for lua 5.1 was removed and the license
text was added to the readme in 5.2.

[Peter: swap 5.1/5.2 logic]
Signed-off-by: Frank Hunleth <fhunleth@troodon-software.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

qwt: bump to version 6.1.0

Signed-off-by: Frank Hunleth <fhunleth@troodon-software.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/mpd: bump to version 0.18.14

Signed-off-by: Jörg Krause <jkrause@posteo.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

alsa-lib: use correct Kconfig symbol for mmu

It is BR2_USE_MMU, not BR2_MMU

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

libffi and python: need threads support

libffi depends on pthreads.

Python depends on libffi (it can provide a builtin libffi, but also depends on
pthreads). Thus this patch also disable Python support if toolchain is compiled
w/o treads support.

Fixes http://sysmic.org/~jezz/results/204099dd:

../src/closures.c:119:21: fatal error: pthread.h: No such file or directory

[Peter: fix comment dependencies, add python3]
Signed-off-by: Jérôme Pouiller <jezz@sysmic.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

directfb: allow SVG support

DirectFB SVG support rely on Cairo and Cairo depends on DirectFB.

Since Cairo DirectFB backend is marked as experimental, this patch choose to
disable this backend when DirectFB SVG support is enable.

In add, this patch may solve some tricky situations between DirectFB and Cairo
(see http://sysmic.org/~jezz/results/c2f11fc0) and improve build reproducibility.

[Peter: propagate atomics dependency, needs libsvg-cairo, add cairo includes]
Signed-off-by: Jérôme Pouiller <jezz@sysmic.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

sdl: remove dga support in no MMU

DGA support use fork. This prevent executables to link with libSDL.so.

Fixes:
  http://sysmic.org/~jezz/results/5e50c243 (package sdl_mixer)
  http://sysmic.org/~jezz/results/734afef8 (package sdl_net)
  http://sysmic.org/~jezz/results/e2310eb9 (package sdl_image)
  http://sysmic.org/~jezz/results/e2cdcc9a (package sdl_sound)

Signed-off-by: Jérôme Pouiller <jezz@sysmic.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

ltp-testsuite: depends on MMU

Fixes http://sysmic.org/~jezz/results/dfc51d2b:

.../ltp-testsuite-20140115/lib/tests/tst_checkpoint_child.c:40: undefined reference to `_fork'
collect2: ld returned 1 exit status
make[3]: *** [tst_checkpoint_child] Error 1
make[3]: *** Waiting for unfinished jobs....
/tmp/ccF9CQQa.o: In function `main':
.../ltp-testsuite-20140115/lib/tests/tst_checkpoint_parent.c:40: undefined reference to `_fork'
collect2: ld returned 1 exit status
make[3]: *** [tst_checkpoint_parent] Error 1
/tmp/ccwCGpYd.o: In function `main':
.../ltp-testsuite-20140115/lib/tests/tst_checkpoint_parent_exits.c:40: undefined reference to `_fork'
collect2: ld returned 1 exit status
make[3]: *** [tst_checkpoint_parent_exits] Error 1
/tmp/cc8TiB6d.o: In function `main':
.../ltp-testsuite-20140115/lib/tests/tst_process_state.c:43: undefined reference to `_fork'
collect2: ld returned 1 exit status
make[3]: *** [tst_process_state] Error 1
/tmp/ccrf7Gjb.o: In function `main':
.../ltp-testsuite-20140115/lib/tests/tst_checkpoint_child_exits.c:40: undefined reference to `_fork'
collect2: ld returned 1 exit status

[Peter: don't mention mmu dependency in comment]
Signed-off-by: Jérôme Pouiller <jezz@sysmic.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

xfsprogs: depends on MMU

Fixes http://sysmic.org/~jezz/results/17f32103:

xfs_fsr.o: In function `fsrallfs':
/home/jezz/br-continuous/cfgs/45-bf609-bfin-uclibc-nommu-ext/build/xfsprogs-3.1.11/fsr/xfs_fsr.c:617: undefined reference to `_fork'
collect2: ld returned 1 exit status

madvise.o: In function `madvise_f':
/home/jezz/br-continuous/cfgs/45-bf609-bfin-uclibc-nommu-ext/build/xfsprogs-3.1.11/io/madvise.c:111: undefined reference to `_madvise'
mincore.o: In function `mincore_f':
/home/jezz/br-continuous/cfgs/45-bf609-bfin-uclibc-nommu-ext/build/xfsprogs-3.1.11/io/mincore.c:77: undefined reference to `_mincore'
collect2: ld returned 1 exit status

[Peter: don't mention mmu dependency in comment]
Signed-off-by: Jérôme Pouiller <jezz@sysmic.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

xterm: depends on MMU

Fixes http://sysmic.org/~jezz/results/53a729c3:

button.o: In function `_executeCommand':
./button.c:(.text+0x21be): undefined reference to `_fork'
main.o: In function `_get_pty':
./main.c:(.text+0xab0): undefined reference to `_getpt'
main.o: In function `_spawnXTerm':
./main.c:(.text+0x1030): undefined reference to `_fork'
misc.o: In function `_creat_as':
./misc.c:(.text+0x3a56): undefined reference to `_fork'
print.o: In function `_charToPrinter':
./print.c:(.text+0x1ee): undefined reference to `_fork'

Signed-off-by: Jérôme Pouiller <jezz@sysmic.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

xapp_xdm: depends on MMU

Fixes http://sysmic.org/~jezz/results/e891eb7e:

dm.o: In function `_StartDisplay':
dm.c:(.text+0xd6e): undefined reference to `_fork'
server.o: In function `_StartServerOnce':
server.c:(.text+0x266): undefined reference to `_fork'
session.o: In function `_runAndWait':
session.c:(.text+0x67a): undefined reference to `_fork'
session.o: In function `_ManageSession':
session.c:(.text+0xa2a): undefined reference to `_fork'
collect2: ld returned 1 exit status

Signed-off-by: Jérôme Pouiller <jezz@sysmic.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

xapp_xmh: depends on MMU

Fixes http://sysmic.org/~jezz/results/1a0bc183:

command.o: In function `__DoCommandToFileOrPipe':
command.c:(.text+0x35e): undefined reference to `_fork'
collect2: ld returned 1 exit status

Signed-off-by: Jérôme Pouiller <jezz@sysmic.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

xapp_xsm: depends on MMU

Fixes: http://sysmic.org/~jezz/results/455f2788:
remote.o: In function `_remote_start':
remote.c:(.text+0x74): undefined reference to `_fork'
restart.o: In function `_Clone':
restart.c:(.text+0x5a8): undefined reference to `_fork'
restart.o: In function `_Restart':
restart.c:(.text+0xa44): undefined reference to `_fork'
collect2: ld returned 1 exit status

Signed-off-by: Jérôme Pouiller <jezz@sysmic.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>