pkg-infra: add hook to log timing of steps

The timing information is stored in the file $(O)/build-time.log

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Reviewed-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Reviewed-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

pkg-infra: introduce pre/post-step hooks

This hooks will let us instrument the build process in many ways:
  - log current step to see what broke
  - time each step to see what is worth optimising
  - sanity-check installed files (rpath, overwritten files...)
  - call user-provided script
  - ...

The steps are coarse-grain, and all have a 'start' and a 'end' hooks.
Here is the list of available steps (8 total):
  - extract
  - patch
  - configure
  - build
  - install-host
  - install-staging
  - install-image
  - install-target

The download, clean and uninstall steps are not instrumented on purpose.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Reviewed-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Reviewed-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

post-{build, images} hooks: export BUILD_DIR too

Also export BUILD_DIR for post-{build,images} hooks, so they do have
a place to store generated files.

Note: this will be more einteresting for the instrumentation of steps,
to come in a later patch.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Reviewed-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Reviewed-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

grep: bump to version 2.15

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

gnutls: bump to version 3.2.6

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

cifs-utils: bump to version 6.2

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

squid: bump to version 3.3.10

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

kickoff 2014.02 development cycle

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

Update for 2013.11-rc1

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

legal info: fix saving of host package licenses

Due to some tricky make behavior, the license texts of host packages that
did not provide an explicit HOST_FOO_LICENSE_FILES definition was not saved.
The problem is that it is not straightforward to use a variable
defined/updated inside an evaluated block as input to a foreach statement.
If you try to use $(FOO) then only the original value of FOO is used for
foreach, any update inside the block is ignored. However, if you use
$$(FOO), the entire contents of FOO (typically a list of items) is passed
as one item to foreach, thus causing just one iteration instead of several.

>From Arnout Vandecapelle's explanation:
Any variable referenced with a single $ inside the inner-generic-package
macro is expanded before the resulting contents are eval'ed. Therefore, it
is not possible to refer to variables defined by the inner-generic-package
macro from within a single-$ function call.

To fix the problem, one should defer the evaluation of the entire block
using double dollar signs.

Additionally, a few empty lines have been added to the legal-info-foo block
for clarity.

Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Acked-by: Luca Ceresoli <luca@lucaceresoli.net>
Tested-by: Luca Ceresoli <luca@lucaceresoli.net>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

tstools: perform shared library linking step using gcc rather than ld

Fixes http://autobuild.buildroot.net/results/c45/c45bbdaff6cff56b5646add38032f29bb5520f16

Using ld directly breaks with certain toolchain configurations.

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

qt5base: add missing hunk in posix_fallocate patch

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

qt5base: detect posix_fallocate() at configure time

Fixes http://autobuild.buildroot.net/results/555278227680e90737e6dbe0a6dcc035507150aa/

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

samba: security bump to version 3.6.20

Fixes CVE-2013-4475 and CVE-2013-4476.

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

CHANGES: update with recent changes

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

docs/news.html: announce dev meeting report

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

docs/about.html: we support internal (e)glibc toolchains as well now

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

gcc: don't patch gcc if there are no patches to apply

While the generic package handler checks for a directory with patches
before starting apply-patches.sh, this is not the case for gcc: the
script is called, even if there is no directory with patches. This results
into a build failure, as apply-patches exits with error code 1 if the
directory doesn't exist.

Signed-off-by: Mischa Jonker <mjonker@synopsys.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

toolchain-external: make sure (e)glibc isn't chosen when BR2_PREFER_STATIC_LIB=y

(e)glibc doesn't support a fully statically linked userspace. Even a
basic program such as Busybox fails to do authentication due to glibc
loading some libraries dynamically. Therefore, we disable the
possibility of using a (e)glibc toolchain when
BR2_PREFER_STATIC_LIB=y.

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

toolchain-buildroot: make sure glibc isn't chosen when BR2_PREFER_STATIC_LIB=y

(e)glibc doesn't support a fully statically linked userspace. Even a
basic program such as Busybox fails to do authentication due to glibc
loading some libraries dynamically. Therefore, we disable the
possibility of building a glibc toolchain when
BR2_PREFER_STATIC_LIB=y.

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

toolchain-buildroot: glibc and eglibc support is no longer experimental

The glibc and eglibc support has been introduced since a little bit of
time now, I believe we can remove the "experimental" statement next to
it.

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

icu: Doesn't work on ARC yet

icu depends on __sync_sub_and_fetch and other atomic primitives that
don't exist in the ARC toolchain yet.

[Peter: adjust beecrypt/php comment dependency, don't mention atomic builtins]
Signed-off-by: Mischa Jonker <mjonker@synopsys.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

Config.in files: whitespace cleanup

This patch fixes the following whitespace problems in Config.in files:
- trailing whitespace
- spaces instead of tabs for indentation
- help text not indented with tab + 2 spaces

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

liblog4c-localtime: fix linking error without pthread

Fixes http://autobuild.buildroot.net/results/34f/34f238cdb1b1b4be1a11143b5e4a17c78c2fc289/

The rollingfile functionality only gets built if pthread support is
available, but a call to these functions from log4c_fini() was outside
the #if WITH_ROLLINGFILE conditional, causing linker errors when the
library is used.

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

libqmi: fix autobuild failures

The libqmi build specifies -Werror in the CFLAGS by default. This causes
build failures such as the following.

  http://autobuild.buildroot.net/results/070/070449299aa9025c4bc6769524a6e22d66c92a65/

The -Werror flag can be disabled using a configuration switch.

[Peter: use =no instead and add a comment explaining why]
Signed-off-by: Simon Dawson <spdawson@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

ethtool: bump version to 3.12.1

3.12 release notes:

* Fix: Remove alternate method to check for VLAN tag offload on Linux
  < 2.6.37 (-k/-K options)
* Fix: Hide state of VLAN tag offload and LRO if the kernel is too old
  for us to reliably detect them (-k option)
* Feature: Add register dump support for Solarflare SFC9100 family
  (-d option)

3.12.1 release notes:

This includes a couple of changes that should have gone into 3.12.

* Fix: Memory corruption when applying external calibration to
  SFF-8472 module diagnostics (-m option)
* Feature: Add Intel 82599 and x540 DCB registers to dump
  (-d option)

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

docs/manual: Add version number to perl in prerequisites

Building host-bison needs perl 5.8.7+, as it uses the "-f" option
for site customization scripts. This feature was added in 5.8.7.

Signed-off-by: Mischa Jonker <mjonker@synopsys.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

pciutils: bump version

I have just released version 3.2.1 of the pciutils. It's purely
a maintenance release with a couple of minor bug fixes and minor
improvements. Thus quoth the changelog:

2013-11-10  Martin Mares <mj@ucw.cz>

* Released as 3.2.1.

* CardBus bridge capabilities are displayed.

* PCIe L1 PM substates are decoded.

* Various bugs were fixed in decoding of PCIe capabilities.

* The sysfs back-end does not spit out unnecessary warnings when
  empty slots report only a partial device address. This actually
  happens on IBM pSeries.

* Updated pci.ids to the today's snapshot of the database.

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

log4cxx: needs threads

Fixes build failures such as the following.

  http://autobuild.buildroot.net/results/7b73901d9e8b5098cd96b54d292bb3ffffb3fe37/

Signed-off-by: Simon Dawson <spdawson@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gst-plugins-ugly: bump version

Bugfix release.

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

gst1-plugins-bad: bump version

Bugfix release. Patch is now upstream.

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

gst1-plugins-good: bump version

Bugfix release. Patch is now upstream.

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

gst1-plugins-base: bump version

Bugfix release.

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

gstreamer1: bump version

Bugfix release.

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

iozone: disable for uClibc 0.9.31

uClibc 0.9.31.x does not include the pthread_setaffinity_np GNU extension.

Fixes build failures such as the following.

  http://autobuild.buildroot.net/results/e21526c0c18733310d3fb322d79d5fff3a71af51/

Signed-off-by: Simon Dawson <spdawson@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

imagemagick: bump version

The -3 version is no longer available at the upstream URL.

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

cppdb: fix comment dependency

As noticed by Thomas De Schampheleir.

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

cppdb: new package

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

xmlstarlet: bump to version 1.5.0

Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

xmlstarlet: add license information

Signed-off-by: Axel Lin <axel.lin@ingics.com>
Acked-by: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

system: move init system option above /dev management.

The option to choose init system was below the one for /dev management.
As Systemd forces the use of udev, it is logical to swap them.

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

package/rpi-userland: bump version

Bump for a few more options (eg. setting the palette in 4- or 8bpp,
and rendering camera preview with GL).

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

package/rpi-firmware: bump version

Bump bootloader and GPU firmware for misc fixes:
  - cec fixes
  - DTS buffering
  - others (pause_burst_frames enabled by default. WTF is that?)

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

p910nd: new package

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

openpowerlink: bump to version 1.8.4

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

libmodbus: bump version and change download location

The license of libmodbus is LGPL v2.1 or later and the
licence of programs in tests directory is GPL v3. So
specify only LGPL v2.1.

Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

hplip: bump to version 3.13.10

Sections of the hplip-fix-make.patch patch file were removed, because they
no longer appear in the Makefile.am file.

Signed-off-by: Olivier Schonken <olivier.schonken@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

iptables: add support for connlabel match and nfnl_osf

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

ipset: bump to version 6.20.1

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

imagemagick: bump to version 6.8.7-3

Signed-off-by: Jerzy Grzegorek <jerzy.grzegorek@trzebnica.net>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

util-linux: disable fallocate for avr32

The fallocate syscall is not implemented in the avr32 toolchain.

Fixes build failures such as the following.

  http://autobuild.buildroot.net/results/bc4/bc41a3fea20181526eb247ac910244aa2aa4c4c0/

Signed-off-by: Simon Dawson <spdawson@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

uboot-tools: factor out common mkimage infrastructure

Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

rootfs-cpio: add U-Boot image support

Adds U-Boot image support for cpio root filesystems. This allows you to
use the bootm command in U-Boot to load the rootfs. It makes it possible to
verify the CRC of the initramfs before booting the kernel.

[Spenser: wrote first version of the patch.]
Signed-off-by: Spenser Gilliland <spenser@gillilanding.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

rootfs-ext2: symlink with the compressed rootfs

Previously, a symlink was created to the uncompressed filesystem,
which made it a bit useless in case compression was chosen.

Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

rootfs-common: refactor the common compression commands

This makes the compression extension available in a variable, so it
can be used by the fs-specific commands. In this patch, it is used
by iso9660. Following patches show more use cases.

Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

rootfs-cpio: whitespace cleanup in Config.in

Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

rootfs-common: remove the now unused ROOTFS_<PKG>_POST_GEN_HOOKS

Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

rootfs-ext2: make the symlink as a _POST_TARGET

This will allow us to remove the unused ROOTFS_$(FSTYPE)_POST_GEN_HOOKS.

Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

toolchain-external: factorize regular expressions

Based on a suggestion from Peter, this commit factorizes the logic and
regular expressions that are used to find the sysroot and libdir for a
given compiler. It reduces a bit the duplication of code, and
centralizes the most bizarre part of this logic in one place.

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

toolchain-external: update comments

[Peter: drop extra # as pointed out by Baruch Siach]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

toolchain-external: fix the SYSROOT_DIR mangling logic

In a1d94aaa3a21911 ('toolchain-external: add support for musl C
library'), we made the following change to the SYSROOT_DIR mangling
logic:

-       SYSROOT_DIR=`echo $${LIBC_A_LOCATION} | sed -r -e 's:usr/lib(32|64)?/(.*/)?libc\.a::'` ; \
+       SYSROOT_DIR=`echo $${LIBC_A_LOCATION} | sed -r -e 's:(usr/)?lib(32|64)?/(.*/)?libc\.a::'` ; \

This was needed to accomodate for musl based toolchains that don't
have libc.a in usr/lib/..., but directory in lib/... Basically, the
change makes the usr/ at the beginning optional.

However, with the very permissive (.*) matching in the middle of the
path, the change above had an unexpected consequence: any path contain
'/lib' would be truncated before this lib. As an example, Peter
reported that his builds, running from /var/lib/buildbot/ were no
longer working because the SYSROOT_DIR was decided to be /var instead
of something like
/var/lib/buildbot/buildroot/output/host/opt/ext-toolchain/arm-linux-gnueabihf/libc/.

So, this commit changes (again!) this regexp by changing (.*) to
([^/]*), the idea being that it will match only *one* path
component. Note that this intermediate (.*) directory was added in
e6e60becb008 ('external-toolchain: add support for Linaro 2012.01') to
accomodate for Linaro toolchains that have a subdirectory in their
sysroot named after the target tuple:

$ ./output/host/opt/ext-toolchain/bin/arm-linux-gnueabihf-gcc -print-file-name=libc.a
/home/thomas/projets/buildroot/output/host/opt/ext-toolchain/bin/../arm-linux-gnueabihf/libc/usr/lib/arm-linux-gnueabihf/libc.a

In addition to this, this commit also makes sure that the change
making usr/ optional is properly reported on all the instances of this
regular expression.

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

infra: clean up 'Patching' message

The 'Patching' message in the generic infrastructure prints not only the
package name, but also a reference to the assumed package directory, based
on FOO_DIR_PREFIX/FOO_RAWNAME. This doesn't really add value, as the name
of the package is already apparent from the message and its location should
be obvious. Hence, this patch simply reduces the print to "Patching".

Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

infra: remove incorrect default for FOO_DIR_PREFIX

Variable FOO_DIR_PREFIX is populated from pkgparentdir by the various
package infrastructures. However, if that would be empty (which in fact is
the case for the linux package), FOO_DIR_PREFIX would be set to
'$(TOP_SRCDIR)/package'.
Not only does this make no sense (LINUX_DIR_PREFIX becomes /package/linux,
and for all other packages pkgparentdir is not-empty anyway), but it is also
using a non-existing variable TOP_SRCDIR.
This patch therefore removes the incorrect default.

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

manual: add some info on the POST_RSYNC hook

Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Acked-by: "Samuel Martin" <s.martin49@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

manual: split info on hooks to a separate section/file

Split out the information on hooks to a separate section (and source file).
Not only because the hooks are useful for all infrastructures (and thus
don't really fit specifically in the generic infrastructure section), but
also for clarity when the info on hooks will be expanded in later patches.

Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Acked-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

infra: centralize rsync exclude list for VCS files

Buildroot has three places where rsync is used:
1. to copy the target skeleton
2. to copy the rootfs overlay(s)
3. to copy overridden package sources

In all of these cases, we want to exclude version control files by default.
Place 1 and 2 used an identical set of explicit --exclude options, while
place 3 used the option --cvs-exclude. This last option, however, not only
excludes version control files, but also binary files (.o, .so) and any file
or directory named 'core' (a problem for the linux kernel that has several
directories with this name). Moreover, the exact list of excluded files when
using --cvs-exclude depends on the version of rsync.

This patch creates one global variable RSYNC_VCS_EXCLUSIONS that can be used
by the various rsync commands. It excludes the version control files of
svn, git, hg, cvs and bzr.

Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Acked-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

infra: Add POST_RSYNC_HOOKS support

One of the use cases is for the 'local packages' to restore
the SCM info.  Some packages use this information to generate
version info during build time.  In this case, the local package
can have this hook to restore it by symbolic link for example.

[Thomas: update commit title]
Signed-off-by: Tzu-Jung Lee <tjlee@ambarella.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Acked-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

Avoid toolchain download when it is preinstalled

For configurations using a toolchain that is preinstalled on
the host, <pkg>_SITE and <pkg>_SOURCE variables must be kept
empty to avoid downloading any toolchain package.
The actual implementation has been proposed by Thomas Petazzoni.

Signed-off-by: GONZALEZ Laurent <br2@gezedo.com>
Reviewed-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Tested-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

Config.in files: add missing dependencies to toolchain option comments

When a package A depends on config option B and toolchain option C, then
the comment that is given when C is not fulfilled should also depend on B.
For example:

config BR2_PACKAGE_A
depends on BR2_B
depends on BR2_LARGEFILE
depends on BR2_WCHAR

comment "A needs a toolchain w/ largefile, wchar"
depends on !BR2_LARGEFILE || !BR2_WCHAR

This comment should actually be:

comment "A needs a toolchain w/ largefile, wchar"
depends on BR2_B
depends on !BR2_LARGEFILE || !BR2_WCHAR

or if possible (typically when B is a package config option declared in that
same Config.in file):

if BR2_B

comment "A needs a toolchain w/ largefile, wchar"
depends on !BR2_LARGEFILE || !BR2_WCHAR

[other config options depending on B]

endif

Otherwise, the comment would be visible even though the other dependencies
are not met.

This patch adds such missing dependencies, and changes existing such
dependencies from
  depends on BR2_BASE_DEP && !BR2_TOOLCHAIN_USES_GLIBC
to
  depends on BR2_BASE_DEP
  depends on !BR2_TOOLCHAIN_USES_GLIBC
so that (positive) base dependencies are separate from the (negative)
toolchain dependencies. This strategy makes it easier to write such comments
(because one can simply copy the base dependency from the actual package
config option), but also avoids complex and long boolean expressions.

Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
 (untested)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

manual: clarify that Config.in comments should have correct dependencies

Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Acked-by: "Samuel Martin" <s.martin49@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

qt5webkit: fix build with bison-3.0

- Add host-bison dependency
- Fetch patch from upstream fixing build with bison-3.0
  (already included in the next 5.2 release)

Fixes:
http://autobuild.buildroot.net/results/283/2831c0d859b035f7d2786f51885833a711b46b80/build-end.log

Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

qt5base: fix rpi/egl support (vc_dispmanx_element_change_attributes proto mismatch)

Patch from upstream already included in the next 5.2 release

Fixes:
http://autobuild.buildroot.net/results/347/347577bf1dee0fec3302a45f278eb253118a5b6f/build-end.log

Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

xorg: needs thread support

libpthread-stubs, libX11, and a few more xorg packages use threads. Since
almost all xorg packages depend on libX11 directory or indirectly, and
since the remaining ones are pretty useless on their own, just require
threads for xorg as a whole.

The thread dependency is kept in libpthread-stubs, because that package
will move out of the x11r7 directory later (pending patch by Spenser
Gilliland), so the dependency on threads will be required then.

Fixes:
  http://autobuild.buildroot.org/results/609/6099baac4bb469ae18aab6512233db25183eaabd/

[Arnout: disable all of xorg, correct comment]
Signed-off-by: Jerzy Grzegorek <jerzy.grzegorek@trzebnica.net>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

xlib_libpthread-stubs: needs -pthread when linking statically

Fixes e.g. http://autobuild.buildroot.net/results/1eaa694263b74313b0c90c510abcb11d490a3773/

Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

Generation of locales: made call to tr more robust and added LOWERCASE macro

When calling 'tr' without quoting braces, bash can make really weird things
if there are existing 'single-letter-named' directories
eg:
thierry@thierry-desktop:~$ echo AAA | tr [A-Z] [a-z]
aaa
thierry@thierry-desktop:~$ mkdir m
thierry@thierry-desktop:~$ echo AAA | tr [A-Z] [a-z]
AAA

The (quick) analysis is that the callee (tr) argvs then
contain 'm' thus the translation does not work

Using quotes works around it:
thierry@thierry-desktop:~$ echo AAA | tr '[A-Z]' '[a-z]'
aaa

Signed-off-by: Thierry Bultel <thierry.bultel@wanadoo.fr>
Reviewed-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

zmqpp: fix build failure

The zmqpp client needs a little help to link against libpthread. Fix build
failures such as the following.

  http://autobuild.buildroot.net/results/d0a/d0a740d4d5414844ca7f2de266fc92223d3d20c8

Signed-off-by: Simon Dawson <spdawson@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libmicrohttpd: disable epoll support on avr32

As it isn't available.

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

uclibc: fix MIPS variables

The BR2_UCLIBC_MIPS_ISA hidden Config.in variable defines which uClibc
config option should be enabled for a given MIPS architecture
variant. Therefore, using lower case names doesn't work: they should
be upper case, to match uClibc config option names. This commit makes
sure the mips32, mips32r2 and mips64 builds select the appropriate
uClibc configuration.

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

freetype: add libpng to freetype-config and freetype2.pc

Add libpng libraries to freetype-config and freetype2.pc when they're
needed to avoid build breakage for other packages.

Patch in a different form is already upstream.

Fixing configure & friends is not so good since autoreconf busts things
up because of the odd way things are done (upstream fixed too it seems).

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

configs: rename rpi_defconfig for consistency

We name all of our boards and defconfig files based on the boards
fullname, not a nickname or a shortname.

'rpi' is short for Raspberry Pi, so name all our Raspberry Pi ressource
with 'raspberrypi' instead of 'rpi'.

This should also help Buildroot-newcomers to recognise Raspberry Pi
related files (defconfig and board doc).

Signed-off-by: Mathieu Benoit <mathieu.benoit@savoirfairelinux.com>
[yann.morin.1998@free.fr: reverse the rename]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Ryan Barnett <rjbarnet@rockwellcollins.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

openssh: bump version

From the announcement:

This release fixes a security bug:

 * sshd(8): fix a memory corruption problem triggered during rekeying
   when an AES-GCM cipher is selected. Full details of the vulnerability
   are available at: http://www.openssh.com/txt/gcmrekey.adv

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

libqrencode: select libpng only when tools are selected

As stated on the libqrencode web site, PNG (or SDL) is only
needed for the test and utility programs; the library itself
has no dependencies.

While we are here, remove spaces in the ifeq clause, to match
the examples in the buildroot manual.

Signed-off-by: Danomi Manchego <danomimanchego123@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

Config.in.legacy: select BR2_LEGACY for module-init-tools legacy option

Commit 94c72087f3 ("module-init-tools: remove package") removed the
module-init-tools package and therefore added the corresponding config
option in Config.in.legacy. However, the commit forgot to add the
"select BR2_LEGACY" that ensures the build cannot proceed until the
user has acknowledged the change (i.e kmod replacing
module-init-tools).

This commit adds this missing "select BR2_LEGACY".

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

qt: remove the BR2_PACKAGE_QT_JAVASCRIPTCORE option

The BR2_PACKAGE_QT_JAVASCRIPTCORE option was available to force the
activation or disabling of the JIT compiler in the Qt Javascript
interpreter. However, the JIT compiler is not available for all
architectures, so forcing its activation does not always
work. Moreover, Qt knows by itself for which architectures JIT support
is possible, and will automatically enable it if possible.

Therefore, this option was in fact useless, and causing build problems
when enabled on architectures for which the JIT support was not
available. This commit removes this option and there is no
replacement: Qt will enable JIT at compile time when possible.

Fixes:

  http://autobuild.buildroot.org/results/aae/aaeb82753b7654eeca679ded5d0211ceebda3ea2/build-end.log
  http://autobuild.buildroot.org/results/367/3670e4f03ff0ce114c90bd7139243d82c427b52a/build-end.log

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

uclibc-snapshot: enable option UCLIBC_HAS_OBSTACK

Prior to uClibc commit f143f920694c, uClibc used to always include the
gnu obstack extension, which is used and required by many tools, such as
binutils.

Because of the change mentioned above, obstack is now optional, and got
disabled per default, so enable it in the uClibc snapshot configuration file.

Signed-off-by: Chris Zankel <chris@zankel.net>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

avr32: fix uclibc build using recent kernel headers

Using kernel headers newer than 3.6.x, uclibc 0.9.31.1 fails to build:

  In file included from output/host/usr/avr32-buildroot-linux-uclibc/sysroot/usr/include/linux/rtnetlink.h:6,
                 from libc/inet/netlinkaccess.h:34,
                 from libc/inet/if_index.c:36:
  output/host/usr/avr32-buildroot-linux-uclibc/sysroot/usr/include/linux/if_link.h:314: error: expected specifier-qualifier-list before '__be16'
  make[1]: *** [libc/inet/if_index.os] Error 1
  make[1]: Leaving directory `output/build/uclibc-0.9.31.1'
  make: *** [output/build/uclibc-0.9.31.1/.stamp_built] Error 2

This patch adjusts the system type definitions in the netlinkaccess.h
header, updating the types to match those used in uClibc 0.9.33.2.

Signed-off-by: Simon Dawson <spdawson@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gutenprint: fix download url

Signed-off-by: Arnaud Aujon <arnaud.aujon@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

openpgm: disable for AVR32, not its Buildroot-built toolchain

OpenPGM builds incorrectly on AVR32 with gcc-4.2.2-avr32-2.1.5.
Since it is presumed to be the only GCC version used by Buildroot
for AVR32, this patch disables openpgm for all cases when AVR32
is selected as the target architecture, including when a toolchain
is downloaded or preinstalled (this is what Buildroot autobuilders
do).

Signed-off-by: Alexander Lukichev <alexander.lukichev@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

dhcpcd: bump to version 6.1.0

Now has proper support for nommu and non-IPv6.

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

mplayer: disable for anything lower than ARMv5

Disable mplayer for anything lower than a v5 ARM core.
The support is broken and it's kind of pointless anyway. Fixes:
http://autobuild.buildroot.net/results/8f946e28b11b2ab1ba0d9688286c665488de0486/

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

trivial: pcmanfm: fix typo in config help

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

uclibc: reinstate support for version 0.9.31

This patch reinstates support for uClibc version 0.9.31, which was removed
from Buildroot in commit 8abb5b33c1aae035cf48b1c81104b433ff884c64

Signed-off-by: Simon Dawson <spdawson@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

openpgm: disable on AVR32

openpgm doesn't build correctly on AVR32 using
gcc-4.2.2-avr32-2.1.5 toolchain: it is configured to call
intrinsic atomic functions not provided by the toolchain,
so they are propagated as unresolved external symbols in the
built openpgm libraries. This breaks programs that try to link
openpgm, because they do not know where to get those either. For
instance, it breaks building zeromq tests when PGM support is
selected.

This commit disables openpgm on AVR32 due to apparent absence of
interest in this package on that architecture and it breaking too
many test builds.

Fixes http://autobuild.buildroot.net/results/5a3261109ea63ba17375003eabd8b5d88757865f/
(at least)

Signed-off-by: Alexander Lukichev <alexander.lukichev@gmail.com>
Reviewed-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

Makefile: fix out-of-tree builds with multiple targets with 'all'

For out-of-tree builds, this use-case fails to build:
    $ make clean all

This is because 'all' is filtered-out in the Makefile wrapper, since
the wrapper itself has a 'all' target.

The 'all' target is just the usual naming for the default target in a
Makefile. In fact, the first target is the default one, so we can name
it whatever we want.

Rename the Makefile wrapper 'all' target to avoid name-clashing.

Fixes #6644.

Reported-by: Ryan Barnett <rjbarnet@rockwellcollins.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Ryan Barnett <rjbarnet@rockwellcollins.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Tested-by: Ryan Barnett <rjbarnet@rockwellcollins.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

rpi: improve documentation

Signed-off-by: Mathieu Benoit <mathieu.benoit@savoirfairelinux.com>
[yann.morin.1998@free.fr: further improvements for readability and
    completeness]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/nut: new package

'nut' are the Network UPS Tools.

[Note: original patch from Dallas, but completely revamped
 by Yann, so nothing remains from the original patch, but two
 variable names.]

Signed-off-by: Dallas Clement <dallas.a.clement@gmail.com>
[yann.morin.1998@free.fr: refresh patch, use latest version,
    add license, fix commit message, add dependencies, fix
    runtime path, fix libs-config, move to 'system utils']
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libevas: disable on avr32

Neither epoll_create1 nor inotify_init1 is available on avr32. Fixes build
failures such as the following.

  http://autobuild.buildroot.net/results/4d435a5fc608936362d605aca696c01023be9723

Signed-off-by: Simon Dawson <spdawson@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

connman: disable on avr32

The inotify_init1 syscall is not available on avr32. Fixes build failures
such as the following.

  http://autobuild.buildroot.net/results/0d8/0d8e7bd8bd3f227fabfb0d5feb59a5d316026d3f

Signed-off-by: Simon Dawson <spdawson@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

avr32: pass target arch to gcc

As suggested by Alexander Lukichev and Thomas Petazzoni on the mailing list,
only one of the two avr32 microarchitectures is relevant for Buildroot:
avr32 Linux implies the avr32b microarchitecure, as used in the ap7000.

Signed-off-by: Simon Dawson <spdawson@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

aircrack-ng: add security patch for CVE-2010-1159

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