rsyslog: bump version to 8.19.0

And drop patch now upstream. Rsyslog now uses libfastjson instead of json-c,
so adjust dependencies.

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

libfastjson: new package

Libfastjson is a fork of json-c, and a dependency of newer versions of
rsyslog.

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

package/luaposix: add .hash file

Similar to jquery-keyboard, it seems the luaposix tarball on the autobuilder is
corrupt.  Fix it by adding a .hash file so it falls back to our s.b.o mirror.

Fixes:
http://autobuild.buildroot.net/results/426/426f85021498611e43cdc2af693099b0687de52c

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/erlang-rebar: add .hash file

Similar to jquery-keyboard, it seems the erlang-rebar tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to our
s.b.o mirror.

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/erlang-p1-zlib: add .hash file

Similar to jquery-keyboard, it seems the erlang-p1-zlib tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to our
s.b.o mirror.

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/erlang-p1-yaml: add .hash file

Similar to jquery-keyboard, it seems the erlang-p1-yaml tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to our
s.b.o mirror.

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/erlang-p1-xml: add .hash file

Similar to jquery-keyboard, it seems the erlang-p1-xml tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to our
s.b.o mirror.

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/erlang-p1-utils: add .hash file

Similar to jquery-keyboard, it seems the erlang-p1-utils tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to our
s.b.o mirror.

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/erlang-p1-tls: add .hash file

Similar to jquery-keyboard, it seems the erlang-p1-tls tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to our
s.b.o mirror.

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/erlang-p1-stun: add .hash file

Similar to jquery-keyboard, it seems the erlang-p1-stun tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to our
s.b.o mirror.

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/erlang-p1-stringprep: add .hash file

Similar to jquery-keyboard, it seems the erlang-p1-stringprep tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to our
s.b.o mirror.

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/erlang-p1-sip: add .hash file

Similar to jquery-keyboard, it seems the erlang-p1-sip tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to our
s.b.o mirror.

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/erlang-p1-iconv: add .hash file

Similar to jquery-keyboard, it seems the erlang-p1-iconv tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to our
s.b.o mirror.

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/erlang-p1-cache-tab: add .hash file

Similar to jquery-keyboard, it seems the erlang-p1-cache-tab tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to our
s.b.o mirror.

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/erlang-lager: add .hash file

Similar to jquery-keyboard, it seems the erlang-lager tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to our
s.b.o mirror.

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/erlang-goldrush: add .hash file

Similar to jquery-keyboard, it seems the erlang-goldrush tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to our
s.b.o mirror.

Fixes:
http://autobuild.buildroot.org/results/074/07463f3478f334ab9c7dda2dbb2a02b4f29faca1

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

linux: allow the selection of the architecture's default configuration

To configure the Linux kernel, we currently provide two options:

 1. Passing a defconfig name (for example "multi_v7"), to which we append
    "_defconfig" to run "make multi_v7_defconfig".

 2. Passing a path to a custom configuration file.

Unfortunately, those two possibilities do not allow to configure the
kernel when you want to use the default configuration built into the
kernel for a given architecture. For example, on ARM64, there is a
single defconfig simply called "defconfig", which you can load by
running "make defconfig".

Using the mechanism (1) above doesn't work because we append
"_defconfig" automatically.

One solution would be to change (1) and require the user to enter the
full defconfig named (i.e "multi_v7_defconfig" instead of "multi_v7"),
but we would break all existing Buildroot configurations.

So instead, we add a third option, which simply tells Buildroot to use
the default configuration for the selected architecture. In this case,
Buildroot will configure the kernel by running "make defconfig".

Cc: Naresh Kamboju <naresh.kamboju@linaro.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

elfutils: put all the "lfs mode" stuff together

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

elfutils: specify the real reason for needing autoreconf

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

elfutils: bump version to 0.166

- Remove 0001-argp-support.patch since it's already included upstream:
  https://git.fedorahosted.org/cgit/elfutils.git/commit/?id=1ab3c2befeedf5bd891cfbe17cfef79c06e5079a

- Tweak patches for this new version:
  - 0002-disable-progs.patch
  - 0003-fts.patch
  - 0004-disable-po.patch
  - 0005-really-make-werror-conditional-to-build-werror.patch

- Remove 0007-Allow-disabling-symbol-versioning-at-configure-t since
  it's already included upstream:
  https://git.fedorahosted.org/cgit/elfutils.git/commit/?id=bafacacaf7659a4933604662daba26a480b29a8d

- Remove portability patch since it's no longer needed. I have asked
  Mark Wielaard regarding this topic given the portability patch looks
  been discontinued since 0.163 and he replied the following:
  - "I believe it is no longer needed. At least not on any platform I
     am actively maintaining. See also this discussion from last year:
     https://lists.fedorahosted.org/pipermail/elfutils-devel/2015-May/004825.html"

- Add zlib dependency since it's now mandatory:
  configure: error: zlib not found but is required

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

libevdev: bump version to 1.5.2

For details see [1].

[1] https://lists.freedesktop.org/archives/input-tools/2016-June/001307.html

Signed-off-by: Peter Seiderer <ps.report@gmx.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

lvm2: bump version to 2.02.156

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

liblog4c-localtime: add .hash file

Fixes:
http://autobuild.buildroot.org/results/e54/e54afe2f5d27540d69e2b9c3b963e23fac5bb4f7/
http://autobuild.buildroot.org/results/e57/e57da53c15f4844b34353154720820aacc9e437a/

Similar to jquery-keyboard, it seems the liblog4c-localtime tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to
our s.b.o mirror.

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

libstrophe: add .hash file

Fixes:
http://autobuild.buildroot.org/results/847/847813211e29bdf8b986696ba5fd5d561e66ca2c/
http://autobuild.buildroot.org/results/b42/b420d52c49ab228f9599f9e03bb55ad3e4dd0fda/

Similar to jquery-keyboard, it seems the libstrophe tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to
our s.b.o mirror.

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

wine: Bump to version 1.8.3

Bumping to stable version 1.8.3

Signed-off-by: André Hentschel <nerv@dawncrow.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libiqrf: add .hash file

Fixes:
http://autobuild.buildroot.org/results/e7a/e7adbfcc963ee3a76368b0f4af7ca63e69995546/

Similar to jquery-keyboard, it seems the libiqrf tarball on the autobuilder
is corrupt.  Fix it by adding a .hash file so it falls back to our s.b.o
mirror.

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

assimp: add .hash file

Fixes:
http://autobuild.buildroot.org/results/f64/f6452c4b0a237219b6ae28c8bb3cf665ad40139f/
http://autobuild.buildroot.org/results/e9a/e9aed367394c6a4ac7b834139b5601e844ba4648/

Similar to jquery-keyboard, it seems the assimp tarball on the autobuilder
is corrupt.  Fix it by adding a .hash file so it falls back to our s.b.o
mirror.

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

python-pyinotify: add .hash file

Fixes:
http://autobuild.buildroot.org/results/35f/35f4c094a3d58b95ed990b57d91dd75f63245cc6/
http://autobuild.buildroot.org/results/0f7/0f7d666e307fc4f665bbe61cdce6c3d6838e73fa/

Similar to jquery-keyboard, it seems the python-pyinotify tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to
our s.b.o mirror.

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

freerdp: add .hash file

Fixes:
http://autobuild.buildroot.org/results/118/1188a06b389072e101061f163f1b7f9fc3ecd092/
http://autobuild.buildroot.org/results/62b/62b2dbeadcb93f821a27d0ff6078b38bdbc44970/

Similar to jquery-keyboard, it seems the freerdp tarball on the autobuilder
is corrupt.  Fix it by adding a .hash file so it falls back to our s.b.o
mirror.

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

pps-tools: add .hash file

Fixes:
http://autobuild.buildroot.org/results/006/00648b4ca73116933a44cccef3dbefbd16766246/

Similar to jquery-keyboard, it seems the pps-tools tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to
our s.b.o mirror.

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

package/bullet: add .hash file

Similar to jquery-keyboard, it seems the bullet tarball on the autobuilder
is corrupt.  Fix it by adding a .hash file so it falls back to our s.b.o
mirror.

Fixes:
http://autobuild.buildroot.net/results/39b/39bb344e3631e5f8872336cef981db5c6ec382c1

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

libdri2: add .hash file

Fixes:
http://autobuild.buildroot.org/results/be2/be2d170dd74f6839fc15b625ee7a24acc74116f6/
http://autobuild.buildroot.org/results/98f/98f9efc226743100ac0411d797bbc59f42f345d0/

Similar to jquery-keyboard, it seems the libdri2 tarball on the autobuilder
is corrupt.  Fix it by adding a .hash file so it falls back to our s.b.o
mirror.

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

tinymembench: add .hash file

Fixes:
http://autobuild.buildroot.org/results/bad/bad9b90db7e7ae352455205747a08e75e8e83376/

Similar to jquery-keyboard, it seems the tinymembench tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to
our s.b.o mirror.

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

libenca: add .hash file

Fixes:
http://autobuild.buildroot.org/results/dfc/dfca1c20a34397bb65d65c3307d3e0d16d7d30ba/
http://autobuild.buildroot.org/results/39c/39c0016972f3894b757f272435c27a38ac075fcb/

Similar to jquery-keyboard, it seems the libenca tarball on the autobuilder
is corrupt.  Fix it by adding a .hash file so it falls back to our s.b.o
mirror.

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

glm: add .hash file

Fixes:
http://autobuild.buildroot.org/results/2c3/2c308c9f63b86f75b4ce8293f8bd1211cb9f4819/
http://autobuild.buildroot.org/results/ff5/ff5058cb59282a8746396f21d700bbaefc859a2d/

Similar to jquery-keyboard, it seems the glm tarball on the autobuilder is
corrupt.  Fix it by adding a .hash file so it falls back to our s.b.o
mirror.

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

librtlsdr: add .hash file

Fixes:
http://autobuild.buildroot.org/results/2a0/2a0e3d64aa1a9a8aaf019c411377b2f4e3c51aae/
http://autobuild.buildroot.org/results/c81/c814311099ff930c961e637b843a8591e8826b75/

Similar to jquery-keyboard, it seems the librtlsdr tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to
our s.b.o mirror.

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

input-event-daemon: add .hash file

Fixes:
http://autobuild.buildroot.org/results/3e1/3e19c1a6bd534c6c5a0be4e68fc272564685e8d8/
http://autobuild.buildroot.org/results/85f/85f7d67664f2806fb86c24f2e87bd72094e7976b/

Similar to jquery-keyboard, it seems the input-event-daemon tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to
our s.b.o mirror.

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

libsha1: add .hash file

Fixes:
http://autobuild.buildroot.org/results/e6d/e6d539a2c47428d1af2529069297af1279c8345f/

Similar to jquery-keyboard, it seems the libsha1 tarball on the autobuilder
is corrupt.  Fix it by adding a .hash file so it falls back to our s.b.o
mirror.

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

gst1-imx: needs shared libraries

Fixes:
http://autobuild.buildroot.org/results/2d1/2d1b5077a7551cebe4b961d6ce1cf844ad83fb8c/

Otherwise waf configure fails with:
Checking for library dl                                      : not found
The configuration failed

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

luaexpatutils: add .hash file

Fixes:
http://autobuild.buildroot.org/results/4f6/4f6439f421809090c874e24055348e3f30b2e2b0/
http://autobuild.buildroot.org/results/bd4/bd4a1db0136ee65f325d1fd9826dedb23986134d/

Similar to jquery-keyboard, it seems the luaexpatutiles tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to
our s.b.o mirror.

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

faketime: add .hash file

Fixes:
http://autobuild.buildroot.net/results/872/8724d7335a699e162aee5f25d8eb2e71310b5c55/
http://autobuild.buildroot.net/results/cdb/cdb6a55efa0a3d86e213aba55ddaeb52d24cf1c6/
http://autobuild.buildroot.net/results/a5f/a5f12f44a2b5d758aed5e4666897cb6eedfc1444/

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

package/smack: add .hash file

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

package/freeswitch: Disable mod_imagick, broken with imagemagick7

Problem sent upstream: https://freeswitch.org/jira/browse/FS-9272

Fixes
http://autobuild.buildroot.net/results/e20/e20c0b9a32f6bef5d43369437895058d44d0df58/
http://autobuild.buildroot.net/results/aa4/aa4d681de85497c2cdd2b0488d9b4569160b5f02/

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

python-crossbar: bump to version 0.14.0

As python-msgpack dependency was replaced with python-u-msgpack,
that is a pure Python implementation, BR2_TOOLCHAIN_HAS_SYNC_4
was removed. BR2_INSTALL_LIBSTDCPP remains because of python-pyasn.

Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

python-sdnotify: bump to version 0.3.1

LICENSE.txt is now part of the PyPi's archive.

Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

log4cpp: needs threads

Fixes:
http://autobuild.buildroot.net/results/cf9/cf942e4b8a5259414a4dcea6fa4b75183435394a/

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

sam-ba: bump to version 2.16 and add hash file

Update also the file name since the pattern has changed, and add a hash
file.

Signed-off-by: Ludovic Desroches <ludovic.desroches@atmel.com>
[Thomas: add hash file.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

libllcp: add .hash file

Fixes:
http://autobuild.buildroot.org/results/d15/d157e77cf948d036a9e36e67534f7b442c0e8bf4/

Similar to jquery-keyboard, it seems the libllcp tarball on the autobuilder
is corrupt.  Fix it by adding a .hash file so it falls back to our s.b.o
mirror.

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

fcgiwrap: add .hash file

Fixes:
http://autobuild.buildroot.org/results/b62/b6200a14da6445eb68915d52211035e326ec82f0/

Similar to jquery-keyboard, it seems the fcgiwrap tarball on the autobuilder
is corrupt.  Fix it by adding a .hash file so it falls back to our s.b.o
mirror.

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

jack2: add .hash file

Fixes:
http://autobuild.buildroot.org/results/cc3/cc3590d9c602b38701398786741fceb612445b92/
http://autobuild.buildroot.org/results/2a6/2a691d01d8c91aa1058960f5176ad611b3549f36/

Similar to jquery-keyboard, it seems the jack2 tarball on the autobuilder
is corrupt.  Fix it by adding a .hash file so it falls back to our s.b.o
mirror.

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

libseccomp: add .hash file

Fixes:
http://autobuild.buildroot.org/results/07f/07fecbce8ca4b0a69e86b93021df80dbc2c45497/

Similar to jquery-keyboard, it seems the libseccomp tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to
our s.b.o mirror.

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

gst1-plugins-bad: disable autoreconf

Autoreconf is no longer required in gst1-plugins-bad and it is not
done for any other gst package either. With autoreconf disabled,
gettextize and a gettext makefile patch are no longer required
and can be safely removed.

Signed-off-by: Marcin Nowakowski <marcin.nowakowski@imgtec.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

logrotate: bump to version 3.9.2

Aside from the version number increase, the package has now moved to
GitHub.

Signed-off-by: James Knight <james.d.knight@live.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

google-material-design-icons: bump to version 2.2.3

Aside from the version number increase, a newly added category `places`
is available; appending to the current icon list.

Signed-off-by: James Knight <james.d.knight@live.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

fb-test-app: add .hash file

Fixes:
http://autobuild.buildroot.org/results/9aa/9aa6818d67d9cfd0bfdf14945476c7783cc05aad/

Similar to jquery-keyboard, it seems the fb-test-app tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to
our s.b.o mirror.

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

cppzmq: add .hash file

Fixes:
http://autobuild.buildroot.org/results/425/425ce97de58c53c4f56d43654c0ec8a7cd7c6430/

Similar to jquery-keyboard, it seems the cppzmq tarball on the autobuilder
is corrupt.  Fix it by adding a .hash file so it falls back to our s.b.o
mirror.

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

json-c: add .hash file

Fixes:
http://autobuild.buildroot.org/results/38f/38fd47a81cc2a6fc515158692af4e2b62f4a965f/
http://autobuild.buildroot.org/results/ecc/ecc4ca7fdda173faf042dfe991073d8c5e9af4d6/

Similar to jquery-keyboard, it seems the json-c tarball on the autobuilder
is corrupt.  Fix it by adding a .hash file so it falls back to our s.b.o
mirror.

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

package/e2fsprogs: fix fsck when using systemd

Commit a93fc34 modified e2fsprogs to symlink /usr/sbin/fsck to /sbin/fsck
to keep systemd happy.  However, later on commit c5bd8af6 has forced
/usr/sbin to be a symlink to /sbin when using systemd.  This means fsck
gets destroyed as a symlink to itself when using systemd or when setting
BR2_ROOTFS_MERGED_USR.

Since commit c5bd8af6 obsoletes commit a93fc34 revert the changes from
commit a93fc34 to solve the issue.

Signed-off-by: Martin Bark <martin@barkynet.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

configs/toradex_apalis_imx6_defconfig: new board

Signed-off-by: Sergio Prado <sergio.prado@e-labworks.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/systemd: Optionally disable hwdb

By default hwdb files are installed in /usr/lib/udev/hwdb.d
Currently they take about ~5MB. After first boot, systemd-hwdb tool
parses them and creates a binary in /etc/udev/hwdb.bin, which takes
another ~6.5MB.

In case of initramfs images, hwdb parsing is done every boot. This adds
about 2-3 seconds to boot time in case of rootfs on SDHC class 10
card (benchmarked with am335x board). NAND boot takes even longer.

Add config option to disable hwdb in systemd, for smaller images and
better boot performance.

Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

opencv: add .hash file

Fixes:
http://autobuild.buildroot.org/results/e06/e06f0e6edd5d8013ff5e12e2b7253e99a1d1fa2e/

Similar to jquery-keyboard, it seems the opencv tarball on the autobuilder
is corrupt.  Fix it by adding a .hash file so it falls back to our s.b.o
mirror.

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

imagemagick: add explicit lzma (xz) support

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

imagemagick: add explicit pango support

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

imagemagick: bump to version 7.0.2-0

Add new disables for libraries that aren't supported in buildroot and
sort them for ease of maintenance.

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

gdb: do not allow selecting the simulator on ARC

Forbid BR2_PACKAGE_HOST_GDB_SIM selection for the ARC architecture,
since the simulator is not supported on ARC.

Fixes:

 http://autobuild.buildroot.net/results/4a1cd9a0beaff34830fd0f0ec47c30eda0a42d54/

Signed-off-by: Vlad Zakharov <vzakhar@synopsys.com>
[Thomas: tweak commit log.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

e2tools: add .hash file

Fixes:
http://autobuild.buildroot.org/results/008/008df073bfa99f93e27d977ef1b64b79ef8b34c4/
http://autobuild.buildroot.org/results/28f/28f0e730768bfaf9577069e4ab18b5a3f3e595a8/

Similar to jquery-keyboard, it seems the e2tools tarball on the autobuilder
is corrupt.  Fix it by adding a .hash file so it falls back to our s.b.o
mirror.

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

snowball-init: add .hash file

Fixes:
http://autobuild.buildroot.org/results/fba/fba69e25de9463282094bcd9b18b63ba6de0a892/

Similar to jquery-keyboard, it seems the snowball-init tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to
our s.b.o mirror.

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

tzdata: bump to version 2016e

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

zic: bump to version 2016e

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

grub2: really disable liblzma

There is no option --enable-liblzma=no in grub2's configure script, so
the only way to disable liblzma support is to pass
ac_cv_lib_lzma_lzma_code=no.

Signed-off-by: Gilles Chanteperdrix <gilles.chanteperdrix@xenomai.org>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
[Thomas: expand commit log, as suggested by Yann E. Morin.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

dawgdic: add .hash file

Fixes:
http://autobuild.buildroot.org/results/66b/66b2989c9441f7fab446d2cc60f047fbed954f25/
http://autobuild.buildroot.org/results/fd6/fd69070ad15792d5a2eea62da6bf22cef78d6f30/

Similar to jquery-keyboard, it seems the dawdgic tarball on the autobuilder
is corrupt.  Fix it by adding a .hash file so it falls back to our s.b.o
mirror.

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

sl: add .hash file

Fixes:
http://autobuild.buildroot.org/results/22e/22e78cfb88f2fc6f6f8b7a0afafbb5eb101bb982/

Similar to jquery-keyboard, it seems the sl tarball on the autobuilder is
corrupt.  Fix it by adding a .hash file so it falls back to our s.b.o
mirror.

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

jsmin: add .hash file

Fixes:
http://autobuild.buildroot.org/results/c05/c05a9da77c4d8ad1e64e029c9e2fd65ebf35772d/
http://autobuild.buildroot.org/results/6ac/6acb5d5c56ece6cefdd135bdb90be035a3200e0a/

Similar to jquery-keyboard, it seems the jsmin tarball on the autobuilder is
corrupt.  Fix it by adding a .hash file so it falls back to our s.b.o
mirror.

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

imx-kobs: bump version to 5.4

Changelog:
- Check raw mode flag for backward compatbility
- Fix membership name in bch geometry structure.
- Fix the previous implementation that may ruin the BBM
- Adapt raw page accesses to match the new raw_read/write implementation

The old raw access implementation (in GPMI driver) was considering that
data and OOB data were separated in their respective regions (the data
area and the OOB area of the page), which is not true.
They are actually interleaved this way:

METADATA + ((DATA + ECCBYTES) * N)

The new raw access implementation (in the GPMI driver) is hiding this
weird
layout to MTD users by exposing a more common layout:

DATA + METADATA + (N * ECCBYTES)

Here METADATA + (N * ECCBYTES) are exposed as if they were stored in the
OOB area.

Unfortunately kobs-ng rely on this weird layout when accessing the NAND
in raw mode.

This patch is based on the Yocto equivalent:
https://github.com/Freescale/meta-fsl-arm/commit/a7545b1

Signed-off-by: Andrew Webster <awebster@arcx.com>
Reviewed-by: Gary Bisson <gary.bisson@boundarydevices.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

imx-lib: bump version to 5.4

Changelog:
- Remove platform check in hdmi-cec and imx-lib - simplify to support builds
on multiple machines
- For ipu use the dummy defaul object to fix build issues on non-ipu hardware

This patch is based on the Yocto equivalent:
https://github.com/Freescale/meta-fsl-arm/commit/7946533

Implicitly tested through gstreamer as the
plugins rely on them for vpu decoding, e.g.:
 # gst-launch-0.10 playbin uri=file:///root/tears_of_steel_1080p.webm
 # gst-launch-1.0 playbin uri=file:///root/tears_of_steel_1080p.webm
 # gst-launch-1.0 filesrc location=/root/tears_of_steel_1080p.webm ! \
   matroskademux ! imxvpudec ! imxipuvideosink

Signed-off-by: Andrew Webster <awebster@arcx.com>
Reviewed-by: Gary Bisson <gary.bisson@boundarydevices.com>
Tested-by: Gary Bisson <gary.bisson@boundarydevices.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

libimxvpuapi: bump version to 0.10.2

Changelog:
- fix AUD NAL positioning in h.264 encoder output
  (SPS/PPS/AUD were ordered incorrectly - AUD has to come first, not last)
- fix build error with examples when --enable-static is used
- pass quality factor in simplified JPEG encoder interface correctly
- add functions for querying and setting header data
  useful for modifying headers, like VUI data in the SPS RBSP
- documentation updates

This patch is based on the Yocto equivalent:
https://github.com/Freescale/meta-fsl-arm/commit/c999668

Implicitly tested through gstreamer as the plugin relies on it:
 # gst-launch-1.0 playbin uri=file:///root/tears_of_steel_1080p.webm

Signed-off-by: Andrew Webster <awebster@arcx.com>
Reviewed-by: Gary Bisson <gary.bisson@boundarydevices.com>
Tested-by: Gary Bisson <gary.bisson@boundarydevices.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

imx-vpuwrap: rename from libfslvpuwrap and bump version to 1.0.65

Changelog:
--Update 'nLastFrameEndPosPhy' in seqinit stage to avoid feed
  the same SPS to vpu repeatedly.

This patch is based on the Yocto equivalent:
https://github.com/Freescale/meta-fsl-arm/commit/d6b9490

Implicitly tested through gstreamer as the 0.10 plugin relies on it:
 # gst-launch-0.10 playbin uri=file:///root/tears_of_steel_1080p.webm

Signed-off-by: Andrew Webster <awebster@arcx.com>
Reviewed-by: Gary Bisson <gary.bisson@boundarydevices.com>
Tested-by: Gary Bisson <gary.bisson@boundarydevices.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

imx-parser: rename from libfslparser and bump version to 4.0.9

Changelog:
--Add aac latm parser in parser_libs, support aac latm in mpeg2 parser

This patch is based on the Yocto equivalent:
https://github.com/Freescale/meta-fsl-arm/commit/e2c6f10

Implicitly tested through gstreamer as the 0.10 plugin relies on it:
 # gst-launch-0.10 playbin uri=file:///root/tears_of_steel_1080p.webm

Signed-off-by: Andrew Webster <awebster@arcx.com>
Reviewed-by: Gary Bisson <gary.bisson@boundarydevices.com>
Tested-by: Gary Bisson <gary.bisson@boundarydevices.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

imx-codec: rename from libfslcodec and bump version to 4.0.9

Changelog:
--[aac_dec] refine the code for pure c build
  aac decoder c build need float point support, which use some
  float point function.

This patch is based on the Yocto equivalent:
https://github.com/Freescale/meta-fsl-arm/commit/4b91194

Implicitly tested through gstreamer as the 0.10 plugin relies on it:
 # gst-launch-0.10 playbin uri=file:///root/tears_of_steel_1080p.webm

Signed-off-by: Andrew Webster <awebster@arcx.com>
Reviewed-by: Gary Bisson <gary.bisson@boundarydevices.com>
Tested-by: Gary Bisson <gary.bisson@boundarydevices.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

imx-vpu: bump to version 5.4.33

Changelog:
VPU firmware has been changed to the format of Linux codec driver firmware.
Change VPU library to adapt to firmware format change.

This patch is based on the Yocto equivalent:
https://github.com/Freescale/meta-fsl-arm/commit/f5e1056

Implicitly tested through gstreamer as the plugins rely on it:
 # gst-launch-0.10 playbin uri=file:///root/tears_of_steel_1080p.webm
 # gst-launch-1.0 playbin uri=file:///root/tears_of_steel_1080p.webm
 # gst-launch-1.0 filesrc location=/root/tears_of_steel_1080p.webm ! \
   matroskademux ! imxvpudec ! imxipuvideosink

Signed-off-by: Andrew Webster <awebster@arcx.com>
Reviewed-by: Gary Bisson <gary.bisson@boundarydevices.com>
Tested-by: Gary Bisson <gary.bisson@boundarydevices.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

gst1-imx: bump to version 0.12.1

Changelog:
* compositor: Rename ImxCompositor(Pad) to ImxVideoCompositor(Pad)
  This prevents conflicts if both the NXP plugins and the gstreamer-imx
  are installed in the same rootfs

* Add switches to wscript configuration to enable/disable plugins
  Until now, all plugins were built (unless their dependencies were
  missing). On some i.MX machines, some features are missing. For
  example, the i.MX7 has no VPU. To ensure deterministic builds, it is
  useful to selectively disable plugins in the build.

* ipu+pxp: : Fix deadlocks and incorrect logging in case of init error
  This was causing a freeze if for example one tries to use PxP elements
  on an i.MX6 Dual/Quad

* imxv4l2videosrc fixes and enhancements:
  * fixes to interlacing support (incl. workarounds for analog TV input)
  * remove format restriction in get_caps function
    this removes the limitation on I420 formats
  * simplified determination of fps (necessary due to driver limitations)

* eglvivsink: wayland: Do not commit before buffers are swapped
  This fixes eglSwapBuffer swaps with damaged frames

* Documentation additions
  Adds GPU memory leak notess to the FAQ

* vpu: encoder: Expose motion estimation range as property

* vpu: decoder: Drop output frame if its buffer cannot be allocated

This patch is based on the Yocto equivalent:
https://github.com/Freescale/meta-fsl-arm/commit/cec4c47

Tested with the following commands on i.MX6Q (IPU):
 # gst-launch-1.0 playbin uri=file:///root/tears_of_steel_1080p.webm
 # gst-launch-1.0 filesrc location=/root/tears_of_steel_1080p.webm ! \
   matroskademux ! imxvpudec ! imxipuvideosink
 # gst-launch-1.0 filesrc location=/root/tears_of_steel_1080p.webm ! \
   matroskademux ! imxvpudec ! imxeglvivsink

Signed-off-by: Andrew Webster <awebster@arcx.com>
Reviewed-by: Gary Bisson <gary.bisson@boundarydevices.com>
Tested-by: Gary Bisson <gary.bisson@boundarydevices.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

firmware-imx: bump to version 5.4

Changelog:
- Upgrade to FW v3.1.1_r46072
- v3.1.1_r46070 plus 2 patches 391/392
- FSLARIK-391 [Decoder/H.264] VPU is always busy when decoding an
erroneous H264 bitstream
- FSLARIK-392 [Decoder/VC1] VPU outputs only 1 picture
- Change to the format of Linux coda driver firmware

This patch is based on the Yocto equivalent:
https://github.com/Freescale/meta-fsl-arm/commit/dbb8773

Implicitly tested through gstreamer as the plugin relies on it:
 # gst-launch-1.0 filesrc location=/root/tears_of_steel_1080p.webm ! \
   matroskademux ! imxvpudec ! imxipuvideosink

Signed-off-by: Andrew Webster <awebster@arcx.com>
Reviewed-by: Gary Bisson <gary.bisson@boundarydevices.com>
Tested-by: Gary Bisson <gary.bisson@boundarydevices.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

xdriver_xf86-video-imx-viv: bump version to 5.0.11-p8.4

Remove upstreamed patch.
Add patch from Yocto to not use git for versioning.

This patch is based on the Yocto equivalent:
https://github.com/Freescale/meta-fsl-arm/commit/06a0574
https://github.com/Freescale/meta-fsl-arm/commit/076df90

The following xorg.conf was used in order to force the use of the
vivante module:
https://github.com/Freescale/meta-fsl-arm/blob/master/recipes-graphics/xorg-xserver/xserver-xf86-config/mx6/xorg.conf

Note that the X server must be started with the "noreset" option to
avoid crashes when closing X apps.
Tested with:
 # X -noreset &
 # cd /usr/share/examples/viv_samples/vdk/
 # DISPLAY=:0 ./tutorial7

Signed-off-by: Andrew Webster <awebster@arcx.com>
[Thomas: add upstream reference in the patch, as suggested by Gary.]

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

imx-gpu-viv: bump version to 5.0.11-p8.4

Changelog:
Software floating point is no longer supported.

For a detailed list of changes in this update, see the Graphics section of
the i.MX Linux Release Notes contained in the i.MX Linux BSP documentation
here:

https://www.nxp.com/webapp/Download?colCode=L4.1.15_1.1.0_LINUX_DOCS&location=null&fpsp=1&WT_TYPE=Supporting%20Information&WT_VENDOR=FREESCALE&WT_FILE_FORMAT=gz&WT_ASSET=Documentation&fileExt=.gz&Parent_nodeId=1337637154535695831062&Parent_pageType=product

This patch is based on the Yocto equivalent:
https://github.com/Freescale/meta-fsl-arm/commit/b667ed4
https://github.com/Freescale/meta-fsl-arm/commit/af5b3a6

This package has been tested with both X11 and Framebuffer backends:
 # cd /usr/share/examples/viv_samples/vdk/
 # apitrace trace --api egl ./tutorial7
 # gmem_info
 ... display memory use per PID ...
 # apitrace replay tutorial7.trace
 # eglretrace tutorial7.trace

Signed-off-by: Andrew Webster <awebster@arcx.com>
Tested-by: Gary Bisson <gary.bisson@boundarydevices.com>
Reviewed-by: Gary Bisson <gary.bisson@boundarydevices.com>
[Thomas:
 - use only one comment for the glibc and ARM EABIhf dependency, like we
   do for the mali-t76x package.
 - reword the comment for the EABIhf dependency, as suggested by Gary
 - remove the IMX_GPU_VIV_BASE_VERSION variable, and use directly
   IMX_GPU_VIV_VERSION, as suggested by Gary.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

kernel-module-imx-gpu-viv: bump version to that of the 5.0.11.p8.4 release

This patch is based on the Yocto equivalent:
https://github.com/Freescale/meta-fsl-arm/commit/c9c30c5

This package has been tested with the following commands with the
matching imx-gpu-viv library:
 # modprobe galcore
 # cd /usr/share/examples/viv_samples/vdk/
 # ./tutorial7

Signed-off-by: Andrew Webster <awebster@arcx.com>
Reviewed-by: Gary Bisson <gary.bisson@boundarydevices.com>
Tested-by: Gary Bisson <gary.bisson@boundarydevices.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

nxp defconfigs: bump kernel version to 4.1.15

There is a 1.1.0_ga release, and a 1.1.1_patch release. The
patch release appears to be only for i.MX6UL, so it is bumped
to 1.1.1_patch, while the other are 1.1.0_ga.

Signed-off-by: Andrew Webster <awebster@arcx.com>
Reviewed-by: Gary Bisson <gary.bisson@boundarydevices.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

nxp defconfigs: bump u-boot version to 4.1.15_1.1.0_ga

Signed-off-by: Andrew Webster <awebster@arcx.com>
Reviewed-by: Gary Bisson <gary.bisson@boundarydevices.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

python-itsdangerous: add .hash file

Fixes:
http://autobuild.buildroot.org/results/596/596f5d6eee03ce94adaca34a6f5b61aed7f929ce/
http://autobuild.buildroot.org/results/a4d/a4d836ab3a0b571b6eab232dc66b5b6ff238d440/

Similar to jquery-keyboard, it seems the python-itsdangerous tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to
our s.b.o mirror.

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

sunxi-mali: add .hash file

Fixes:
http://autobuild.buildroot.org/results/3e2/3e26462743269001510ac8a072cd61c86f2a63b3/

Similar to jquery-keyboard, it seems the sunxi-mali tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to
our s.b.o mirror.

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

libtpl: add .hash file

Fixes:
http://autobuild.buildroot.org/results/1af/1af513c5832ad44d348fa99ea2b65f3a8e462249/
http://autobuild.buildroot.org/results/18d/18d492dc166c3d175e00f2d621d66d209854cb93/

Similar to jquery-keyboard, it seems the libtpl tarball on the autobuilder
is corrupt.  Fix it by adding a .hash file so it falls back to our s.b.o
mirror.

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

libatomic_ops: : add .hash file

Fixes:
http://autobuild.buildroot.org/results/a45/a459be0bf948c3d5db6d5ab7cd9a1a89edd692a2/
http://autobuild.buildroot.org/results/866/866b35878b03e0046b18285dcff89e164a658237/

Similar to jquery-keyboard, it seems the libatomic_ops tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to
our s.b.o mirror.

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

mplayer: not available on m68k

Fixes:
http://autobuild.buildroot.org/results/e48/e48db9dc487547f4d6f6102b4bd2512e0281018a/
http://autobuild.buildroot.org/results/cfc/cfc6ad13c7b4fb080ee89ba4acb420ebb38373ea/

From the log:
The architecture of your CPU (m68k) is not supported by this configure script
It seems nobody has ported MPlayer to your OS or CPU type yet.

Error: unsupported architecture m68k

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

package/freeswitch: bump version to 1.6.9

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

python-twisted: bump to version 16.2.0

Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

python-autobahn: bump to version 0.14.1

Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

python-txaio: bump to version 2.5.1

Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

python-sdnotify: new package

Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

cpuload: add .hash file

Fixes:
http://autobuild.buildroot.org/results/e59/e595a49833bf3046af6fd6ef4ddb7443d40771b9/

Similar to jquery-keyboard, it seems the cpuload tarball on the autobuilder
is corrupt.  Fix it by adding a .hash file so it falls back to our s.b.o
mirror.

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

lpc3250loader: add .hash file

Fixes:
http://autobuild.buildroot.org/results/a60/a601f30d7100821ad0e303c6338348751d7aa99b/
http://autobuild.buildroot.org/results/56f/56f4e5f864da7e2176c78426d1d180c37966b181/

Similar to jquery-keyboard, it seems the lpc3250loader tarball on the
autobuilder is corrupt.  Fix it by adding a .hash file so it falls back to
our s.b.o mirror.

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

fbterm: uses fork(), needs mmu

Fixes:
http://autobuild.buildroot.org/results/7d7/7d7081e7ce6e2be71b8df3b6213142c5ceeeccce/
http://autobuild.buildroot.org/results/54c/54cc2f443eddc1459b2006683f001d473ba65924/

And at the same time, cleanup the 'depends on' lines.

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