toolchain-external: Update Sourcery CodeBench MIPS toolchain

Update to the latest update for the 2013.05 toolchain
released on 11 September 2013.

https://sourcery.mentor.com/GNUToolchain/release2554

Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

luacrypto: refactor with autotools

some CMake stuff comes from LuaDist, but CMake is not the primary build infrastructure of LuaCrypto.

Signed-off-by: Francois Perrad <francois.perrad@gadz.org>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

apply-patches.sh: avoid reversed/already applied patches

Disable reversed/already applied patches fallout from commit
5871b791995ebe295db7dca608afe3f293ce8953
Reverse patches are bad, they may unfix things with version bumps and
just sneak under the radar with pure batch mode.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

ltrace: bump to version 0.7.3

Now with proper mips support!

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

aiccu: needs threads

Fixes:
http://autobuild.buildroot.net/results/f6e/f6e1da89b3a0418a50cbfa39fc9373dad3fc968c/

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

olsr: bump to version 0.6.6

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

iproute2: bump to version 3.11.0

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

ethtool: bump to version 3.11

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

ruby: security bump to version 1.9.3-p448

Fixes CVE-2013-4073.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

quagga: security bump to version 0.99.22.3

Fixes CVE-2013-2236.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

dash: bump to version 0.5.7-3

Bump to patchlevel 3 since 2 is gone upstream, fixes when
sources.buildroot.net isn't available:
http://autobuild.buildroot.net/results/5ee/5ee6544a9d08025060cb5806870ae1c3f252cd93/

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

screen: needs mmu

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

time: needs mmu

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

docs: website: update for 2013.08.1

Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

can-utils: bump version

Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

libroxml: bump version to 2.2.3

Signed-off-by: Tristan Lelong <tristan.lelong@blunderer.org>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

Makefile: release: create tarball from current branch, not master

To support stable bugfix releases.

Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

apply-patches.sh: Go back to only warn about unsupported file types

Fixes http://autobuild.buildroot.net/results/868/8687be8ec029486d9c5e2224cde542134f72884b/

The recent (d245fbb41dc: apply-patches.sh: detect missing patches)
change to apply-patches.sh causes a number of regressions with packages
using downloadable tarballs of patches (typically from Debian), as
those contain additional files besides just the patches (ChangeLog's,
debian/rules, ..).

This use case is arguably abusing the _PATCH handling, but it used to
work so people might rely on it so go back to only warn about this
instead of erroring out.

At the same time reword the warning message.

Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

docs/news.html: Announce next Buildroot developer days meeting

Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

docs/news.html: add 2013.08 announcement mail link

Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

bellagio: needs dynamic library support

Fixes:
http://autobuild.buildroot.net/results/76e/76ebe0b9e963367427e23ba963f1bce6fd2725c4/

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

Trivial documentation fix for Signed-off-by

Make consistent the capitalization and punctuation of Signed-of-by examples.

Signed-off-by: Ralph Siemsen <ralphs@netwinder.org>
Acked-by: "Samuel Martin" <s.martin49@gmail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

docs/patch-policy.txt: fix missing word

Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

toolchain: update Linaro AArch64 external toolchains

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

toolchain: update Linaro ARM external toolchains

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

gcc: provide option to enable mudflap support

The mudflap library is only useful if one uses the -fmudflap gcc
option to do more checks on pointers/arrays. This commit adds an
option to enable/disable mudflap support at the gcc level. By default,
it is disabled, which saves a little bit of build time compared to the
default of gcc which consists in enabling mudflap support.

Since mudflap is now disabled by default, and ensured to never be enabled
on platforms where it is not available, some gcc.mk code that was used to
disable mudflap in problematic configurations can be removed.

Whether -fmudflap is used when building is left to the user.

[Peter: tweak commit text, fix white space]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

toolchain: move elf2flt option inclusion to toolchain-buildroot/Config.in

The only remaining thing in toolchain-buildroot/Config.in.2 is the
inclusion of the elf2flt option. It doesn't really make sense to have
a separate Config.in file for that, so let's move this to
toolchain-buildroot/Config.in.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

toolchain: refactor Stack Smashing Protection support

This commit refactors how Stack Smashing Protection support is handled
in Buildroot:

 *) It turns the BR2_TOOLCHAIN_BUILDROOT_USE_SSP option into an option
    that only enables the SSP support in uClibc, when using the internal
    toolchain backend.

 *) It adds an hidden BR2_TOOLCHAIN_HAS_SSP option that gets enabled
    when the toolchain has SSP support. Here we have the usual dance:
    glibc/eglibc in internal/external backend always select this
    option, in the case of uClibc/internal, it gets selected when
    BR2_TOOLCHAIN_BUILDROOT_USE_SSP is enabled, in the case of
    uClibc/external, there is a new configuration option that the user
    must select (or not) depending on whether the toolchain has SSP
    support.

 *) It adds a new options BR2_ENABLE_SSP in the "Build options" menu,
    to enable the usage of SSP support, by adding
    -fstack-protector-all to the CFLAGS.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

gcc: force symlink creation

The HOST_GCC_CONFIGURE_SYMLINK creates a symbolic link
$(@D)/build/configure -> $(@D)/configure for each build step of
gcc. However, it was only using 'ln -s' and not 'ln -sf', which was
causing problems when doing 'make host-gcc-final-reconfigure' for
example, because the configure commands (including this macro) are
being re-executed, but the symbolic link already exists.

Changing this symbolic link creation to 'ln -sf' fixes this problem.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

gcc/gcc-final: use the common HOST_GCC_CONFIGURE_SYMLINK macro

Each stage of the gcc build needs to make a small dance before the
configuration step to create a build sub-directory and a symbolic link
to the configure script. The common gcc.mk had a
HOST_GCC_CONFIGURE_SYMLINK already used by the gcc-initial and
gcc-intermediate steps, but the gcc-final step wasn't using it.

This commit fixes this inconsistency, and therefore removes the
HOST_GCC_FINAL_CONFIGURE_SYMLINK macro.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

toolchain: add support for glibc

The support for eglibc 2.17 was added to the internal toolchain
backend for 2013.08. This commit now adds glibc 2.18 support to the
internal toolchain backend.

Since the building procedure is very similar to the one of eglibc, we
have renamed the 'eglibc' package to 'glibc', and made it capable of
handling either glibc or eglibc.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

gcc: move C++ support option next to Fortran/Objective-C

The option to enable C++ support was still located in
toolchain/toolchain-buildroot/Config.in.2, with misc other toolchain
options. It seems more logical to have this option with the other
options to select the languages supported by the cross-compiler, so we
move it next to the Fortran/Objective-C options in
package/gcc/Config.in.host.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

gcc: fix Objective-C support

This commit fixes the Objective-C support in the gcc package by making
sure that the runtime Objective-C library is properly installed to the
target.

It also takes this opportunity to slightly reword the Config.in option
help text to have a more meaningful explanation.

Finally, it fixes the gcc-final.mk code to use the correct
BR2_INSTALL_OBJC option rather than the non-existing
BR2_GCC_CROSS_OBJC option, to decide whether Objective-C support
should be enabled or not in the cross-compiler.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

gcc: fix Fortran support

This commit fixes the Fortran support in the gcc package by making
sure that the runtime Fortran library is properly installed to the
target.

It also takes this opportunity to slightly reword the Config.in option
help text, to no longer mention gcc < 4.2 since they are no longer
supported in Buildroot, and to have a more meaningful explanation.

Finally, it fixes the gcc-final.mk code to use the correct
BR2_INSTALL_FORTRAN option rather than the non-existing
BR2_GCC_CROSS_FORTRAN option, to decide whether Fortran support should
be enabled or not in the cross-compiler.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

toolchain-internal: skip gcc-intermediate when possible

When NPTL support was introduced, gcc required a three stages build
process. Since gcc 4.7, this is no longer necessary, and it is
possible to get back to a two stages build process. This patch takes
advantage of this, by doing a two stages build process when possible.

We introduce a few hidden kconfig options:

 * BR2_GCC_VERSION_NEEDS_THREE_STAGE_BUILD, which is set by the gcc
   Config.in logic to indicate that the compiler might need a three
   stages build. Currently, all versions prior to 4.7.x are selecting
   this kconfig option.

 * BR2_TOOLCHAIN_LIBC_NEEDS_THREE_STAGE_BUILD, which indicates whether
   the C library might need a three stages build. This is the case for
   eglibc, and uClibc when NPTL is enabled.

 * BR2_TOOLCHAIN_NEEDS_THREE_STAGE_BUILD finally is enabled when both
   of the previous options are enabled. It indicates that a three
   stages build is actually needed.

In addition to those options, the uClibc/gcc build logic is changed to
use only a two stages build process when possible.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

eglibc: libstdc++ is not part of the C library

For some reason, the eglibc.mk file was trying to install libstdc++.so
from eglibc to the target. But the C++ standard library is provided by
GCC, not by the C library.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

package/Makefile.in: fix incorrect comment

The TARGET_{CC,CXX,LD,...} variables no longer contain any --sysroot
option, since we're now using a toolchain wrapper for external
toolchains.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

system: default to devtmpfs for /dev

As discussed on the list, devtmpfs is quite a bit nicer default than
the old static /dev, so change it.

Notice that you NEED to enable CONFIG_DEVTMPFS / CONFIG_DEVTMPFS_MOUNT
in the kernel configuration if you manually build a kernel outside
buildroot, otherwise the rootfs won't work.

Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

getty: add the ability to pass options to getty

Signed-off-by: Assaf Inbal <shmuelzon@gmail.com>
Acked-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

getty: Create specific getty config + cleanups

This commit introduces a specific BR2_TARGET_GENERIC_GETTY configuration flag.
This eliminates the need for checking if BR2_TARGET_GENERIC_GETTY_PORT is an
empty string or not. It also allows hiding various getty options when getty
isn't enabled.

Signed-off-by: Assaf Inbal <shmuelzon@gmail.com>
Acked-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

apply-patches.sh: detect missing patches

The "patch" command returns an error code only if patches fail
to apply. Therefore the pipleline "cat <patchfile> | patch ..."
does not fail, even if <patchfile> is missing. Fix this by
adding an explicit check for patch file existence.

Based on feedback from buildroot mailing list, also change the
existing check for unsupported patch format into a fatal error.

Signed-off-by: Ralph Siemsen <ralphs@netwinder.org>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

Makefile: make $(BUILD_DIR)/.root rule idempotent

The $(BUILD_DIR)/.root rule is executed as part of the 'dirs'
target. The 'dirs' target is re-executed at every execution of 'make
external-deps', and make external-deps explicitly tells make to ignore
targets that have already been made (through the -B option). This
means that the $(BUILD_DIR)/.root rule has to be idempotant, which was
not the case this the introduction of the lib32/lib64 symbolic link.

Running 'make external-deps' three times in a row was sufficient to
trigger an error due to symbolic links being incorrectly created. This
patch fixes that.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

aiccu: new package

Automatic IPv6 Connectivity Configuration Utility for users of a
IPv6 tunnel broker, developed by sixxs.net

[Peter: drop strip patch, pass TARGET_CONFIGURE_OPTS]
Signed-off-by: Michael Rommel <rommel@layer-7.net>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

squid: fix build failure after bump

Add missing patch, fixes:
http://autobuild.buildroot.net/results/f6a/f6a6dd1f3138406d1daa31e5e4ff5d5026b3c069/

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

omniorb: new package

[Peter: fix C++ dependency and trailing spaces, drop ';' from fixup hook]
Signed-off-by: Matt Weber <mlweber1@rockwellcollins.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

linux: bump 3.11.x stable version

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

linux-headers: bump 3.{0, 4, 10, 11}.x stable versions

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

tcping: new package

tcping does a TCP connect to the given ip/port combination. The user can
specify a timeout in seconds. This is useful in shell scripts running in
firewalled environments. Often SYNs are just being dropped by firewalls,
thus connection establishment will be retried several times (for minutes)
until a TCP timeout is reached. With tcping it is possible to check first
if the desired port is reachable and then start connection establishment.

http://www.linuxco.de/tcping/tcping.html

[Peter: wrap help text, pass TARGET_LDFLAGS]
Signed-off-by: Arnaud Rébillout <rebillout@syscom.ch>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

package: remove the empty trailing lines

Signed-off-by: Jerzy Grzegorek <jerzy.grzegorek@trzebnica.net>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

nfs-utils: fix build failure caused by missing include

Fixes http://autobuild.buildroot.net/results/a7fabcf61d56bca4bb005a91f8f367ffc8474ea0/

Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

elf2flt: upgrade to latest version

Upgrade to the latest uclinux CVS snapshot, gives us blackfin support.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

nfacct: add patch for uclinux tuple

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

libnetfilter_queue: add patch for uclinux tuple

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

libnetfilter_cttimeout: add patch for uclinux tuple

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

libnetfilter_cthelper: add patch for uclinux tuple

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

ulogd: needs mmu and shared objects

The plugin infrastructure is based on shared objects so it won't build
for static-only scenarios.
And the daemon uses fork() so MMU is required.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

libnetfilter_log: add patch for uclinux tuple

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

libnetfilter_conntrack: add patch for uclinux tuple

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

libnfnetlink: add patch for uclinux tuple

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

libnetfilter_acct: add patch for uclinux tuple

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

ipset: add patch for uclinux tuple

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

libmnl: add patch for uclinux tuple

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

nommu/flat: build packages with the correct FLAGS

When building for nommu flat targets the CFLAGS/CXXFLAGS/LDFLAGS must be
adjusted accordingly.
For gcc this means passing along -Wl,-elf2flt to signal the linker.
For ld this means -elf2flt.

Also correct the error in STACKSIZE settings from commit 9edf482d which
is setting gcc flags as ld flags and will surely fail (no package uses
it at the moment so it was never seen).

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

toolchain/gcc: add elf2flt to dependencies

Add host-elf2flt to the gcc common dependencies so we get it built after
binutils but before gcc/uclibc since it's required for all packages and
in some uClibc configuration scenarios.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

docs/manual: add CVS support documentation

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

downloads: add basic CVS support

The support is for pserver mode anonymous CVS.
source-check is based on login since many servers don't support or have
ls/rls disabled.

Usage is pretty straightforward.
PKG_SITE defines the site hostname and remote directory.
The module is defined by the bare package name.
Version is date based.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

igmpproxy: new package

[Peter: minor white space fixes]
Signed-off-by: Matt Weber <mlweber1@rockwellcollins.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

ola: new package

[Peter: fixup help text]
Signed-off-by: Dave Skok <blanco.ether@gmail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

libsoc: strip trailing newline

Missing from previous commit.

Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

libsoc: add new package

[Peter: fix help text, doesn't need fork, file header, license tags]
Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

ubifs filesystem image: line up and improve Config.in options

Triggered by the 'some comment required here', this patch updates the Config.in
comments for the ubifs filesystem image options. This includes:
- 'erase block' --> 'eraseblock' (following mtd online documentation)
- remove 'UBI' prefix to make options uniform. Sometimes the option is not a
  property of UBI but rather of the flash. Also, line up the name of the option
  with the help of mkfs.ubifs.
- add help text on all options, referring to the mkfs.ubifs/ubinize options

Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

knock: new package

A port knocking implementation with daemon and user application.

[Peter: drop trailing newline in Config.in]
Signed-off-by: Michael Rommel <rommel@layer-7.net>
Acked-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

ser2net: bump to version 2.9.1

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

usb_modeswitch_data: bump to version 20130807

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

usb_modeswitch: bump to version 2.0.1

Drop libusb-compat since it's no longer required.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

dosfstools: bump to version 3.0.22

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

squid: bump to version 3.3.9

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

libusb: build fix for flat binaries

Fixes:
http://autobuild.buildroot.net/results/4d1/4d1b9c6132e7836bc2ecab4b1483b80d2e042dcd/

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

Makefile: also unexport PKG_CONFIG_LIBDIR

Similar to how we do it for _PATH / _SYSROOT_DIR. Buildroot gets fairly
confused if the user has set up a custom PKG_CONFIG_LIBDIR, so ensure
it gets unset.

Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

linux.mk: target-install: drop extra (invalid) DEPMOD parameter

DEPMOD is already specified in LINUX_MAKE_FLAGS, and is no longer located
in HOST_DIR/usr/sbin after the move to host-kmod, so drop it from here.

Reported-by: Ryan Coe <bluemrp9@gmail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

package: drop unneeded 'call' from {generic,autotools}-package invocation

Reported-by: Jerzy Grzegorek <jerzy.grzegorek@trzebnica.net>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

pkg-stats: update set of files to skip

In BR sub-directory boot/ linux/ and package/ there are a few .mk files which
aren't <package>.mk files. These files shouldn't be taken into account
in package statistics.

Signed-off-by: Jerzy Grzegorek <jerzy.grzegorek@trzebnica.net>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

linux-headers: bump 3.{0, 2, 4, 10}.x stable versions

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

vsftpd: Fix dependency for needs mmu

Commit 2eb995759 "vsftpd: needs mmu" adds "depends on BR2_TOOLCHAIN_HAS_THREADS".

The build error is:
sysutil.c:(.text+0x37ac): undefined reference to `fork'

Thus it should depend on BR2_USE_MMU rather than BR2_TOOLCHAIN_HAS_THREADS.

Cc: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

socketcand: needs mmu

Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

noip: needs mmu

Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

netplug: needs mmu

Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

syslinux: bump to version 4.07

Signed-off-by: Romain Naour <romain.naour@openwide.fr>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

xserver_xorg-server: backport kdrive option parsing fix

Reported-by: Woody Wu <narkewoody@gmail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

barebox: bump to version 2013.09.0

Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

linux: fix depmod location after move to host-kmod

The host-kmod version of depmod gets installed into HOST_DIR/sbin, so
adjust the path we use to refer to it in linux.mk

Reported-by: Ryan Coe <bluemrp9@gmail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

libunwind: needs threads

Fixes:
http://autobuild.buildroot.net/results/39a/39a32d323252e3d986fd9290108db7a496d429d9/

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

libroxml: needs threads

Fixes:
http://autobuild.buildroot.net/results/169/169b607993cd7ae4a8a79c2a00c00f820ec1c747/

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

opus-tools: disable flac when it is not selected

opus-tools expects flac since the 0.1.7 bump unless otherwise told so.
Disable it when it's not selected. Fixes:
http://autobuild.buildroot.net/results/c8c/c8c1a682336c150ee401f76b173944611b96bd4e/

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

erlang: install to staging

This is required to build custom port drivers/NIFs or code that
uses erl_interface.

Signed-off-by: Hans Bulfone <hans@nil.at>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

Config.in: Fixup after BR2_ELF2FLT got renamed to BR2_PACKAGE_HOST_ELF2FLT

The BR2_ELF2FLT option has been renamed to BR2_PACKAGE_HOST_ELF2FLT, thus
use the new name to fix dependency.

Signed-off-by: Axel Lin <axel.lin@ingics.com>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

mpd: remove reference to old multimedia sub-directory

Signed-off-by: Sagaert Johan <sagaert.johan at skynet.be>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

pulseaudio: fix init script installation

Adjust the installation step after adf65110 (Remove multimedia subdirectory).

Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

util-linux: fix build failure with older systems

Adding patch util-linux-002-support-older-machines.patch to fix issues
when building for older systems.

Making all util-linux patches apply cleanly (fuzz fixes)

Fixes http://autobuild.buildroot.org/results/349/349d6ff938b093623618669a1acb390aa2a5fafe
Fixes http://autobuild.buildroot.org/results/57e/57e2c612c46b9e3b5f7bb213fcb3825b23fb7fc7

Signed-off-by: Ryan Barnett <rjbarnet@rockwellcollins.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>

package: don't use .xz format for github

Fixes http://autobuild.buildroot.net/results/955/955ed1af28f6ff9912adddc0342d1f3ca65e5b4d/

Commit 62146ea3ad7 (change package tarball compression to xz whenever possible)
changed a number of packages to use .xz format tarballs. Unfortunately this
doesn't work for packages from github, as it delivers gzip compressed tarballs
even when a .xz is requested.

Fix it by reverting this change for the packages using github.

Reported-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>