openjpeg: refresh patches

The 2.1.2 version bump did not take into account the recently added fixes.
Update the patches to apply based on upstream pull requests #866 and #867.

Cc: Peter Seiderer <ps.report@gmx.net>
Cc: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Acked-by: Peter Seiderer <ps.report@gmx.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/gcc: backport xtensa trap pattern for 5.x

gcc 5.x uses trap pattern extensively to mark locations where program
execution should never get. Default pattern implementation emits a call
to 'abort' function. This function however is not always available, e.g.
it's not available for the libc dynamic linker. Backport implementation
of the trap pattern for xtensa that does not result in a call to
'abort'.

Fixes:
  http://autobuild.buildroot.net/results/0fbcc4475545904bf88f25a3bdf3ee552bf4960e
  http://autobuild.buildroot.net/results/cb7f5bdfac615d774f3bcd1b802f8d74460bc557
  http://autobuild.buildroot.net/results/cf3dc9ae4f74ff28b285f1a862b71d96384d7651

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

dtv-scan-tables: rename file to have only ASCII characters

Since the bump of dtv-scan-tables to version
ceb11833b35f05813b1f0397a60e0f3b99430aab in commit
b1c8794d8ac0eb3895d13ae91d8e912ec469a105, one file contains non-ASCII
characters, which causes encoding issues tvheadend. Since no other
file in the dtv-scan-tables code base contains files with non-ASCII
characters (despite having files named after cities in various
countries that definitely do have non-ASCII characters), we rename
this file so that it is named with only ASCII characters.

This fixes the build of tvheadend, which was failing when the host
Python interpreter was python3, due to a file name encoding issue.

Fixes:

  http://autobuild.buildroot.net/results/1ae8bee297edb089535a2fb6ec724ebf7976888d/
  (tvheadend)

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

openjpeg: security bump to version 2.1.2

See CHANGELOG.md for the full list of fixes, including security issues.

See CVE number lists at [1] and [2].

[1] http://advisories.mageia.org/MGASA-2016-0362.html
[2] https://lists.fedoraproject.org/archives/list/package-announce@lists.fedoraproject.org/thread/HPMDEUIMHTLKMHELDL4F4HZ7X4Y34JEB/

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

toolchain: Bump ARC tools to arc-2016.09-rc1

As described at:
4520524ba055706236db9f00dd79f1b2e2e87fde
this commit continues a series of updates of ARC tools.
This time we're updating tools to arc-2016.09-rc1.

This update contains a lot of important fixes, e.g. it fixes:
http://autobuild.buildroot.net/results/4c7/4c77f33c842b37bf28cb931edf1b290e1bf4d93c//
http://autobuild.buildroot.net/results/902/902729a0b98675ad803939e3ecdcf230065a6012//
and other failures.

Other important change is that we also update gdb. Now we are
using gdb 7.12.

This version of gdb requires C++ toolchain support so we add
corresponding dependency to gdb Config.in file.

Signed-off-by: Vlad Zakharov <vzakhar@synopsys.com>
[Thomas:
 - fix dependency on C++ of gdb, it must use BR2_INSTALL_LIBSTDCPP
 - add comment about the C++ dependency of gdb on ARC.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

configs/snps_axs*: bump U-Boot to 2016.11

With this commit we upgrade U-Boot version to 2016.11 in
"snps_axs101_defconfig" and "snps_axs103_defconfig".

Important fixes: new version fixes U-Boot build for arc700.  Since
gcc-6.x "-marc700" option is no longer supported, "-mcpu=arc700"
should be used instead.

We haven't sent it before as we were waiting for U-Boot 2016.11
release.

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

config: bump linux kernel to 4.8.6 in synopsys defconfigs

With this commit we update ARC defconfigs with the following:

  - "snps_axs101_defconfig", "snps_axs103_defconfig" and
    "snps_hs38_smp_vdk_defconfig":
     - bump linux kernel version to 4.8.6
     - set up host linux headers to 4.8

Signed-off-by: Vlad Zakharov <vzakhar@synopsys.com>
[Thomas: as reported by Alexey, this commit is in fact a bug fix: it
makes the defconfig work properly. Indeed, since the ARC compiler has
moved to gcc 6.x, the ABI has changed, and a Linux 4.8+ kernel is
needed for userspace to work on ARC.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

Update for 2016.11-rc2

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

CHANGES: mention quota package update

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

CHANGES: update for -rc2

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

quota: fix build on musl

This commit adds 3 patches to the quota package, that allows to build
it with musl. One patch was borrowed from OpenEmbedded, the other two
patches are really trivial.

Fixes:

  http://autobuild.buildroot.net/results/6f0ff905251d12e198654b8ffe0ea5c15186371e/

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

openjpeg: fix static linking of liblcms2

Fixes [1]:

  .../host/usr/bfin-buildroot-uclinux-uclibc/sysroot/usr/lib/liblcms2.a(cmsplugin.o): In function `_cmsDeleteContext':
  cmsplugin.c:(.text+0x1c2): undefined reference to `_pthread_mutex_lock'
  cmsplugin.c:(.text+0x1f6): undefined reference to `_pthread_mutex_unlock'

http://autobuild.buildroot.net/results/5ce/5cee20afd8bef5268832cddcb3a5270746be7a57

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

lcms2: fix lcms2.pc.in thread library dependency (for static linking)

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

package/openjpeg: fix static build

This change adds a patch to openjpeg fixing the tiff indirect
dependencies in case of static build.

A similar patch for upstream master has been submitted [1].

Fixes:
  http://autobuild.buildroot.net/results/d0d/d0d22727311d6300e0e400728126170407bfd699/
  and many others...

[1] https://github.com/uclouvain/openjpeg/pull/866

Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Cc: Olivier Schonken <olivier.schonken@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/strace: disable libmpers with musl toolchains

strace bundle some kernel headers to build libmpers, this mixes userspace
headers and kernel headers which break the build with musl.
The stddef.h from gcc is used instead of the one from musl.

Disable libmpers for musl toolchains.

Fixes:
http://autobuild.buildroot.net/results/6f6/6f63fa2537ff316ba259a20966faadcf5ab9deff

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

tinymembench: add patch to fix build on Cortex-M

This commit adds a patch to the tinymembench package that fixes the
build on Thumb-only architectures such as ARM Cortex-M. The patch has
been submitted upstream in
https://github.com/ssvb/tinymembench/pull/13.

Fixes:

  http://autobuild.buildroot.net/results/5f0148cef22b3b9aee691e1702722a0bbc68c8e8/

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

package/weston: fix build without EGL

Fixes:
    http://autobuild.buildroot.org/results/378/378ffddcebd0ab3dabfdb268614d1344a38e8dd1/
    http://autobuild.buildroot.org/results/3c2/3c25333d3c5838bea27540d8ac752c8bdf4db843/
    [...]

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

jasper: disable -pedantic-errors

The -pedantic-errors gcc option turns -pedantic warnings into errors. This
mostly affects older gcc versions that default to the ISO90 C standard. Use
the --disable-strict configure option to remove -pedantic-errors.

Fixes:
http://autobuild.buildroot.net/results/191/191f80779df1a9e6f832106e6c4bdf601e2a9893/
http://autobuild.buildroot.net/results/1fe/1febccc7215814490fa3c776b34bc367363afe39/
http://autobuild.buildroot.net/results/a6f/a6f9bfec3406fc21b130f1669e3534651b9c9596/

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

linux: bump defaul to version 4.8.7

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

linux-headers: bump 3.12.x and 4.{4, 8}.x series

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

package/sane-backends: fix musl build issue

We should include config.h from sanei_backend.h in order to use the
correct if/else HAVE_FOO.

For some reason with Glibc or uClibc there is no problem but with musl
we have the following weird issue:

In file included from epsonds.h:41:0,
                 from epsonds-jpeg.c:18:
../include/sane/sanei_backend.h:99:33: error: expected ';', identifier or '(' before 'int'

../include/sane/sanei_backend.h:99:33: warning: useless type name in empty declaration

That's because HAVE_SIGPROCMASK is not defined although it's correctly
detected by the configure script.

$ grep config.log
config.log:#define HAVE_SIGPROCMASK 1

So, include config.h to avoid to redefine sigset_t.

Fixes:
http://autobuild.buildroot.net/results/9f1/9f1f1cb727b5c5407e69172280a3dee880e55cdf

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Cc: Frank Hunleth <fhunleth@troodon-software.com>
Reviewed-by: Frank Hunleth <fhunleth@troodon-software.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

wget: fix static link with gnutls

When statically linking with gnutls, we get definition clash error for
base64_encode which is also defined by gnutls.

This commit adds patch to rename base64_{encode,decode} defined in wget.

Signed-off-by: Rahul Bedarkar <rahul.bedarkar@imgtec.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

wget: fix ssl detection in static libs configuration

When building wget with openssl in static libs configuration, wget
build system fails detect openssl because it doesn't specify LD flags
for private libs used by openssl. This specifically happens when we
pass --with-libssl-prefix to configure which tries to find ssl using
custom flags. If we don't specify --with-libssl-prefix, it relies on
pkg-config files to detect ssl and it's LD flags which helps with static
linking.

This commit removes --with-libssl-prefix conf opts. Since this case is
similar to gnutls, we remove same conf opts for gnutls as well.

wget can be built with either gnutls or openssl crypto libraries, so
separate optional support for both is not required. This commit also
does minor optimization by checking for either gnutls or openssl while
at it.

Fixes:
  http://autobuild.buildroot.net/results/c6a/c6abdff37b86471cf8b0ceffeff5472042923de0/

Signed-off-by: Rahul Bedarkar <rahul.bedarkar@imgtec.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

go: remove powerpc64 big-endian from supported architectures

cgo currently doesn't properly support powerpc64 big-endian, as noted
in https://github.com/golang/go/issues/13192, and indeed, we have a
large number of build failures of Go packages on this
architecture. This commit therefore disables Go on PowerPC64
big-endian (PowerPC64 little-endian is fine).

Fixes:

  http://autobuild.buildroot.net/results/a6e9bac0a735f48d0ba0af081aeac4ed9fdfaca7/
  (flannel)

  http://autobuild.buildroot.net/results/230f52bc35f437836c7a76d4b58ef454635ee0d3/
  (docker-containerd)

  http://autobuild.buildroot.net/results/77c31d6e8f5efe3e024e27a160cf5d1d1952719e/
  (runc)

  http://autobuild.buildroot.net/results/a87b07417ea8bd81ffe27e5661b4359ddc0149ab/
  (docker-engine)

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

package/kvmtool: fix a musl build issue on ARM

We simply bump the version rather than backport the patch, as there has
been only very few, minor commits in-between.

The kvmtool build with musl still produce some warning but they are
fixed with the current master 0093df80 (kvmtool: Makefile: disable PIE
build for bios and pre_init). So the next version bump should fixes
them.

Fixes:
http://autobuild.buildroot.net/results/8e8/8e81354e82a8c0495e2ca465eb79cb8cf4c9b153

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Cc: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

jasper: security bump to version 1.900.22

Fixes:
CVE-2016-8693: Double free vulnerability in mem_close
CVE-2016-8692: Divide by zero in jpc_dec_process_siz
CVE-2016-8691: Divide by zero in jpc_dec_process_siz
CVE-2016-8690: Null pointer dereference in bmp_getdata triggered by crafted
BMP image
CVE-2016-2089: matrix rows_ NULL pointer dereference in jas_matrix_clip()
CVE-2016-8886: memory allocation failure in jas_malloc
CVE-2016-8887: Null pointer dereference in jp2_colr_destroy
CVE-2016-8884, CVE-2016-8885: Null pointer dereference in bmp_getdata
(incomplete fix for CVE-2016-8690)
CVE-2016-8880: Heap buffer overflow in jpc_dec_cp_setfromcox()
CVE-2016-8881: Heap buffer overflow in jpc_getuint16()
CVE-2016-8882: Null pointer access in jpc_pi_destroy
CVE-2016-8883: Assert in jpc_dec_tiledecode()

Drop upstream patches.

Change SITE to the official download location, since the current one does not
have the updated version. Unfortunately, the official site only offers tar.gz.

Fix license. It is "based on the MIT license", but not exactly the same
(http://www.ece.uvic.ca/~frodo/jasper/; under "Legal Issues").

Drop autoreconf; the autotools version has been updated since commit
324ccec90d (jasper: autoreconf to fix rpath issue) that introduced it.

Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/efl: remove wayland dependency

Wayland support requires DRM support in efl libraries since 1.18 bump,
but this support is not merged yet in Buildroot and waiting for review
[1].

For 2016.11, keep wayland support disabled to avoid build issues.

Fixes:
http://autobuild.buildroot.net/results/162133cbddcca1ea8be660a19771f04296c20fd2

[1] http://patchwork.ozlabs.org/patch/671260/
    http://patchwork.ozlabs.org/patch/671261/

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

package/domoticz: disable for Sourcery CodeBench SH 2012.09

Like for prboom, domoticz trigger an internal compiler error:
[...]/hardware/MySensorsBase.cpp: In member function 'void MySensorsBase::Do_Send_Work()':
[...]/hardware/MySensorsBase.cpp:2190:6: internal compiler error: Segmentation fault

Fixes:
http://autobuild.buildroot.net/results/0ea/0eadf342f86ed71c3a4008f13a4d3e175143460d

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

configs/atmel_sama5d4_xplained: re-add defconfig

This defconfig was mistakenly removed by commit
78ac39cb89a6933c36508f6b10d30150ddb1ba05 due to an improper conflict
detection in "git am". This commit restores it.

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

board/atmel/readme.txt: remove mention of sama5d4ek

The sama5d4ek defconfig has been removed in commit
9cb0b37cbd267ec1c2beea3deeaa86846bc2d037, but the board/atmel/readme.txt
file was not updated accordingly. This commit fixes that.

Reported-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Ludovic Desroches <ludovic.desroches@atmel.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

configs/atmel_sama5d4ek: remove defconfig

The Atmel sama5d4ek board is no longer sold. Since there is the Xplained
board and we won't maintain anymore the EK board, the defconfig can be
removed. In addition, the defconfig was failing to build with gcc 5.x.

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

configs/atmel_sama5d3xek: bump to linux4sam_5.5

Bump to linux4sam_5.5 and remove dtb variants which are no longer
supported. This is needed to fix the build with gcc 5.x.

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

configs: atmel: at91sam9g45m10ek: update defconfig

Bump to a recent version of AT91bootstrap and use mainline version of
U-Boot and Linux. This is needed to fix the build with gcc 5.x.

Signed-off-by: Ludovic Desroches <ludovic.desroches@atmel.com>
[Thomas: add missing BR2_PACKAGE_HOST_LINUX_HEADERS_CUSTOM_4_8=y.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

configs: atmel: at91sam9g20dfc: update defconfig

Bump to a recent version of AT91bootstrap and use mainline version of
U-Boot and Linux. This is needed to fix the build with gcc 5.x.

Signed-off-by: Ludovic Desroches <ludovic.desroches@atmel.com>
[Thomas: add missing BR2_PACKAGE_HOST_LINUX_HEADERS_CUSTOM_4_8=y.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

configs: atmel: at91sam9rlek: update defconfig

Bump to a recent version of AT91bootstrap and use mainline version of
U-Boot and Linux. This is needed to fix the build with gcc 5.x.

Signed-off-by: Ludovic Desroches <ludovic.desroches@atmel.com>
[Thomas: add missing BR2_PACKAGE_HOST_LINUX_HEADERS_CUSTOM_4_8=y.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

configs: atmel: at91sam9260eknf: update defconfig

Bump to a recent version of AT91bootstrap and use mainline version of
U-Boot and Linux. This is needed to fix the build with gcc 5.x.

Signed-off-by: Ludovic Desroches <ludovic.desroches@atmel.com>
[Thomas: add missing BR2_PACKAGE_HOST_LINUX_HEADERS_CUSTOM_4_8=y.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/pseudo: provide better legacy handling for fakeroot

We now have a wrapper that makes pseudo behaves like the fakeroot of the
good ol' days. So the symlink will just magically keep old scripts
working as they did before the switch to pseudo.

However, using fakeroot is deprecated, and we want people to stop using
it altogether and switch to pseudo.

So, make the wrapper recognise how it's called, and if called as
fakeroot, print a warning message.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Gaël PORTAY <gael.portay@savoirfairelinux.com>
Cc: Patrick Keroulas <patrick.keroulas@savoirfairelinux.com>
Cc: Erico Nunes <nunes.erico@gmail.com>
Cc: Julien BOIBESSOT <julien.boibessot@free.fr>
Cc: Arnout Vandecappelle <arnout@mind.be>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

fs: call the pseudo wrapper

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Gaël PORTAY <gael.portay@savoirfairelinux.com>
Cc: Patrick Keroulas <patrick.keroulas@savoirfairelinux.com>
Cc: Erico Nunes <nunes.erico@gmail.com>
Cc: Julien BOIBESSOT <julien.boibessot@free.fr>
Cc: Arnout Vandecappelle <arnout@mind.be>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/pseudo: wrap the real pseudo with a wrapper

Running pseudo is more involved than running fakeroot. In the transition
from using fakeroot, we just did not account for the extra requirements.

First, we explicitly tell pseudo where it is, otherwise it tries to
guess. Its guess is correct, but it prints a warning, which is not nice.

Second, we tell it where to find the passwd and group files in case it
has to emulate access to them. We currently do not use that feature, but
better safe than sorry.

Third, pseudo spawns a background daemon, and talks to it (when fakeroot
would emulate the state all in the current process' state, pseudo uses
the daemon to coordinate the state across multiple processes). We are
not much interested in the daemon lingering around, so we just tell it
to terminate as soon as the last clients quits (this can take up to one
second).

Fourth and last, pseudo always stores its internal database when
exiting, and reloads it when spawned. The database is by default stored
in a sub-directory of the prefix it was installed in, but this is
impractical for us. We want the database to be specific to the one
config dir we are building, so we store the database in a (hidden)
sub-dir of the build dir, thus ensuring it is never shared with another
build. That directory is hidden (starts with a dot) because we consider
that to be our internal state that we do not want to expose to the user.

The wrapper has to be relocatable, so we avoid using hard-coded paths
in there: we derive those paths fom the runtime path of pseudo. However,
the build directory $(BUILD_DIR) is not available in the environment
(we do not export it because it conflicts with some buildsystems).
Instead, we use $(BASE_DIR) which is exported.

Finally, when relocated, the wrapper would not be used in the Buildroot
environment, so may not have access to TARGET_DIR or BASE_DIR, unless
the user sets them. If he does not, we still want the wrapper to be
working (to avoid the warning about the prefix, and to exit the daemon
asap); thus we leave the passwd and localstatedir variable alone if we
don't have what it needs to set them, rather than set them to incorrect
values.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Gaël PORTAY <gael.portay@savoirfairelinux.com>
Cc: Patrick Keroulas <patrick.keroulas@savoirfairelinux.com>
Cc: Erico Nunes <nunes.erico@gmail.com>
Cc: Julien BOIBESSOT <julien.boibessot@free.fr>
Cc: Arnout Vandecappelle <arnout@mind.be>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/pseudo: force rpath and pass our host CFLAGS

pseudo whines when those are not set, and tries to find a suitable
value; this is usually correct but risk an incorrect guess in corner
cases.

Rather than leaving those in guess-mode, just force them to values we
know are correct.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Gaël PORTAY <gael.portay@savoirfairelinux.com>
Cc: Patrick Keroulas <patrick.keroulas@savoirfairelinux.com>
Cc: Erico Nunes <nunes.erico@gmail.com>
Cc: Julien BOIBESSOT <julien.boibessot@free.fr>
Cc: Arnout Vandecappelle <arnout@mind.be>
Reviewed-by: Arnout Vandecappelle <arnout@mind.be>
[Thomas: use double quotes instead of single quotes.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

Makefile: fix stripping of ld.so and libpthread for merged /usr

If 'lib' is a symlink (as is the case when BR2_ROOTFS_MERGED_USR=y),
'find lib' does not return the correct result. So, until now,
libpthread*.so* and ld-*.so* were not stripped when 'lib' was a symlink.

We fix this by using 'find lib/' instead of 'find lib'. For consistency
reason, we also do the same change for the 'find' that removes .a and
.la files.

Signed-off-by: Jérôme Pouiller <jezz@sysmic.org>
[Thomas: slightly improved the commit log.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

DEVELOPERS: remove Gregory Hermant

The Calao Systems company has gone bankrupt in April 2016, and the
domain name no longer exists, so there is no chance to ever contact
Gregory Hermant at a @calao-systems.com address, so let's remove his
entry from the DEVELOPERS file.

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

configs: remove calao*_defconfig

The calao defconfigs no longer build with gcc 5.x, due to Linux kernel
versions used being too old. However, it is unlikely that anyone will
ever update them, since Calao Systems has gone bankrupt in April 2016.

Therefore, let's remove them. If anyone is interested again at some
point, it will be easy to revive them from the Git history.

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

package/guile: dont't use -Os for CS 2014.05 ARM toolchain

With this toolchain, guile tigger an assembler error with -Os which is
probably caused by the binutils version being used:

arm-none-linux-gnueabi-ld --version
GNU ld (Sourcery CodeBench Lite 2014.05-29) 2.24.51.20140217

arm-none-linux-gnueabi-gcc --version
arm-none-linux-gnueabi-gcc (Sourcery CodeBench Lite 2014.05-29) 4.8.3 20140320 (prerelease)

Nowaday, even with a Buildroot internal toolchain using the oldest
Binutils GCC version (2.25.1 and 4.8.6), guile build fine with -Os.

So, force -O2 when this toolchain is used and BR2_OPTIMIZE_S is set.

Fixes:
http://autobuild.buildroot.net/results/976/976b3477407e64c8c2fc2309ed952d0083903a19

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

MIPS: rename M6201 core to M6250

m6201 is the -march option for GCC, but the real core name is
M6250.

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

MIPS: rename M5101 core to M5150

m5101 is the -march option for GCC, but the real core name is M5150.

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

MIPS: remove M5100 core

This is a microcontroller class (MCU) core which is not suitable for
running Linux.

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/pseudo: enforce the host bitness

pseudo can detect the host bitness, but is not sure about it: it checks
what type of file /bin/sh is, using file(1).

However, in some conditions, /bin/sh can be of a different bitness than
the rest of the system (weird, but not impossible), which causes build
issues.

Just enforce the bitness, so that pseudo needs not (wrongly) guess it.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Gaël PORTAY <gael.portay@savoirfairelinux.com>
Cc: Patrick Keroulas <patrick.keroulas@savoirfairelinux.com>
Cc: Erico Nunes <nunes.erico@gmail.com>
Cc: Julien BOIBESSOT <julien.boibessot@free.fr>
Cc: Arnout Vandecappelle <arnout@mind.be>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

php-fpm: remove config comments to fix startup

The comments in the Buildroot-provided php-fpm.conf would produce the
following error when starting php-fpm:

ERROR: [/etc/php-fpm.conf:2] value is NULL for a ZEND_INI_PARSER_ENTRY

Removing the comments fixes the problem.

Signed-off-by: Frank Hunleth <fhunleth@troodon-software.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

DEVELOPERS: add Cyril Bur to PowerPC

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

valgrind: do not use host mpicc

Valgrind may use 'mpicc' (from openmi project) to compile libmpiwrap-*.so.
Without any option, it will detect and use 'mpicc' from host and generate
libmpiwrap-*.so for host:

  $ file target/usr/lib/valgrind/libmpiwrap-arm-linux.so
  target/usr/lib/valgrind/libmpiwrap-arm-linux.so: ELF 64-bit LSB  shared object,
  x86-64, version 1 (SYSV), dynamically linked, not stripped

We simply disable openmi support for valgrind.

Signed-off-by: Jérôme Pouiller <jezz@sysmic.org>
[Thomas: refactor with other unconditional CONF_OPTS.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

php: fix bfin compile issue

Fixes:
  http://autobuild.buildroot.net/results/8326f24511d3b109d22a76660adc017bbcaca197

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

python3: fix double format detection

Python is not able to detect if compiler double representation is
compliant with IEE754:

  checking whether C doubles are little-endian IEEE 754 binary64... no
  checking whether C doubles are big-endian IEEE 754 binary64... no
  checking whether C doubles are ARM mixed-endian IEEE 754 binary64... no

Accordingly 'legacy' mode isused. It is possible to check this at
runtime by check if 'sys.float_repr_style' contains 'short' or
'legacy'. Calculus correctness is not garanteed with 'legacy'.

Problem is better described here:

  http://stackoverflow.com/questions/29920294/what-causes-pythons-float-repr-style-to-use-legacy
  https://bugs.python.org/issue7117

However, all gcc architecture use a representation compliant with
IEE754. So, we can enable it unconditionnaly.

Signed-off-by: Jérôme Pouiller <jezz@sysmic.org>
[Thomas: rework condition to not use strip, as suggested by Baruch.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

python: fix double format detection

Python is not able to detect if compiler double representation is
compliant with IEE754:

  checking whether C doubles are little-endian IEEE 754 binary64... no
  checking whether C doubles are big-endian IEEE 754 binary64... no
  checking whether C doubles are ARM mixed-endian IEEE 754 binary64... no

Accordingly 'legacy' mode isused. It is possible to check this at
runtime by check if 'sys.float_repr_style' contains 'short' or
'legacy'. Calculus correctness is not garanteed with 'legacy'.

Problem is better described here:

  http://stackoverflow.com/questions/29920294/what-causes-pythons-float-repr-style-to-use-legacy
  https://bugs.python.org/issue7117

However, all gcc architecture use a representation compliant with
IEE754. So, we can enable it unconditionnaly.

Signed-off-by: Jérôme Pouiller <jezz@sysmic.org>
[Thomas: adjust condition to avoid usage of qstrip, suggested by Baruch.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

olimex_a20_olinuxino_lime2: bump U-Boot to 2016.05

The patch fixes build by GCC 5 (unsupported by the older U-Boot).

Signed-off-by: Jan Viktorin <viktorin@rehivetech.com>
Reported-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

ci20_defconfig: use XBurst CPU

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

arch/Config.in.mips: add support for XBurst cores

The Ingenic XBurst is a MIPS32R2 microprocessor.

It has a bug in the FPU that can generate incorrect results in certain
cases. The problem shows up when you have several fused madd
instructions in sequence with dependant operands.

Using the -mno-fused-madd option prevents gcc from emitting these
instructions. This patch adds changes to the toolchain wrapper to use
that option.

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

configs: freescale_imx31_3stack: bump kernel version to 4.1.15_2.0.0_ga

Switch to the kernel of release 4.1.15_2.0.0_ga, as it builds properly with
gcc 5.x, which is now the default.

We add a linux config fragment to disable the framebuffer, to repair the
build for imx_v6_v7_defconfig.

Suggested-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Vincent Stehlé <vincent.stehle@laposte.net>
Signed-off-by: Julien Olivain <juju@cotds.org>
Tested-by: Julien Olivain <juju@cotds.org>
Cc: Fabio Estevam <fabio.estevam@nxp.com>
[Thomas: adjust syntax in fragment.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

libxslt: add upstream security fix

Fixes CVE-2016-4738 - heap overread.

https://www.debian.org/security/2016/dsa-3709

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

libxslt: correct homepage link

Link to the actual webpage of libxslt, not to the xslt spec page.

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

mpv: fix build on powerpc64 w/ altivec

Fixes
http://autobuild.buildroot.net/results/77f/77f00633b0b25f001cd4ca54ed3a12aee301c287

Add a patch to work around a name collision between bool and altivec's
bool apparent when using sdl2.

Signed-off-by: Sam Bobroff <sam.bobroff@au1.ibm.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/ccache: bump to version 3.3.3

Note, that version 3.3.3 of ccache fixes a bug [1] where the Linux
kernel image is the same even after changing the source code and
rebuilding it.

[1] https://github.com/ccache/ccache/issues/136

Signed-off-by: Jörg Krause <joerg.krause@embedded.rocks>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

libmpeg2: add patch to fix build on ARMv7-M

This commit adds a patch to libmpeg2 that improves the ARM detection
to not use the optimized ARM code on ARMv7-M architectures.

Fixes:

  http://autobuild.buildroot.net/results/9c44de035ce81c6cc8a92ea8bad4def25cd99cd0

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

assimp: disable package with the ADI Blackfin toolchain

assimp fails to build with an internal compiler error with the ADI
Blackfin toolchain, while it builds fine with the mainline gcc for
Blackfin. So let's disable this package (which has no reverse
dependencies that select it) for the ADI Blackfin toolchain.

Fixes:

  http://autobuild.buildroot.net/results/394e3545a7bbe5d6fbaf4c97f0a3eb51c7d57076

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

ffmpeg: disable on Microblaze

Building ffmpeg on Microblaze triggers the gcc PR71124, which causes
an infinite loop in the compiler. This causes numerous timeouts in the
autobuilders, such as, just over the last two days:

http://autobuild.buildroot.net/results/b47fb2ae810ada475fef215c1efb54a9891fef6f
http://autobuild.buildroot.net/results/a8a1dfb7d4a7a6babe665214030e2245d572d1d3
http://autobuild.buildroot.net/results/248c58cd20686497c4dabc1a19b2c8c7c76b33ab
http://autobuild.buildroot.net/results/3641092f62551bd41c24eafdce46804f9551b761
http://autobuild.buildroot.net/results/f2f65106c3e9dcfd546d5dad7cd0f7cd7145420c
http://autobuild.buildroot.net/results/c45f085ca421b12195fe6a9298b3b3a9cb761360
http://autobuild.buildroot.net/results/349de9b8251b9b72ace595920303e45497833c29

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

net-tools: fix field with incomplete type error

Update 0002 patch to match the one sent as a pull request. The previous
one introduced a new problem that has been addressed in the new version.

Some Linux headers included in certain toolchains may not have an
updated linux/if_tunnel.h which includes linux/ip.h, so we need to
include it unconditionally otherwise linux/if_tunnel.h will use the
struct iphdr before being defined and the compilation will fail in this
way:

In file included from iptunnel.c:33:0:
/usr/include/linux/if_tunnel.h:37:16: error: field 'iph' has incomplete
type
  struct iphdr  iph;
                  ^

Pull request URL:
  https://sourceforge.net/p/net-tools/code/merge-requests/4/

Fixes:
  http://autobuild.buildroot.net/results/34f/34f9b10ef6a613057e8f6457d8a0a2fa142ad123/

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

board/boundarydevices: fix boot partition number in bootscript

It was previously hardcoded to use the 1st partition although the
bootpart variable was used to download kernel/dtb from any partition.

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

DEVELOPERS: add Yegor Yefremov for the ser2net package

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

ser2net: bump to version 3.1

Version 3.1 adds librt to the list of required libraries.

Fixes:
http://autobuild.buildroot.net/results/7f1/7f152e2255f8d2cf9378812998fe9a239319c116

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

configs/qmx6: remove defconfig

The original submitter and follow-up contributors to the qmx6
defconfig no longer have the hardware available or the time/interest
to update this defconfig, which currently fails to build with gcc
5.x. Due to this, this commit removes this defconfig.

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

configs/freescale_p2020ds: remove defconfig

This defconfig no longer builds with gcc 5.x, and the original
submitter, Gustavo, no longer has the hardware to test this
configuration.

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

board/armadeus: bump apf28 kernel version to 4.4.30 (latest LTS)

Signed-off-by: Julien BOIBESSOT <julien.boibessot@armadeus.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

board/armadeus: bump apf51 kernel to 4.4.30 (latest LTS)

This will fix:
http://free-electrons.com/~thomas/pub/buildroot-defconfigs/armadeus_apf51_defconfig.log.FAIL

Signed-off-by: Julien BOIBESSOT <julien.boibessot@armadeus.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/qemu: disable Qemu user-land emulation for musl toolchains

Like for cyclictest (rt-test)[1], linux-user/syscall.c use the
definition of the "struct sigevent" from Glibc/uClibc which doesn't
build with musl libc.

So, disable Qemu user-land emulation for musl toolchains.

Fixes:
http://autobuild.buildroot.net/results/b27/b277e668d9e22b624416bf46f6ccba802acd257e

[1] 43bffa59e7359610c8f00038ddd97795c0ca35e1

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Cc: Francois Perrad <francois.perrad@gadz.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

olimex_a20_olinuxino_lime_mali: bump u-boot to 2016.09.01

U-Boot KConfig allows to take care of CONFIG_OLD_SUNXI_KERNEL_COMPAT
in order to properly start the kernel (Error: unrecognized/unsupported machine ID)

Signed-off-by: Francois Perrad <francois.perrad@gadz.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

configs/mini2440_defconfig: remove defconfig

This defconfig has not been updated in any significant way since 2011,
when Peter moved it to use a mainline kernel. It's now using an
ancient 3.0 kernel, and a vendor-specific U-Boot version that fails to
build with various "undefined reference" errors:

  undefined reference to `show_boot_progress'
  undefined reference to `image_print_contents'
  undefined reference to `wait_ms'

So it's time to get rid of the defconfig for this platform.

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

Revert "package/x265: bump version to 2.1"

This reverts commit 1ba7b6a4f12977aec994c4f38d5be38cfdb27846.

Adding -DCROSS_COMPILE_ARM=ON to _CONF_OPTS will not fix the build:

/home/buildroot/br2/output/host/usr/arm-buildroot-linux-gnueabihf/sysroot/usr/include/gnu/stubs.h:7:29:
  fatal error: gnu/stubs-soft.h: No such file or directory

because cross-compiling x265 is only supported for soft-float abi:
https://bitbucket.org/multicoreware/x265/src/583fc74fc0a29f330187dbd78151c30a3e03d5a7/source/CMakeLists.txt?at=default#CMakeLists.txt-211

Fixes
http://autobuild.buildroot.net/results/6ec/6ec75cc176e05bfc6287dc54fa1a53e0ffc0521d/

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

package/mesa3d: Remove opengl headers files if opengl support is disabled

Mesa3d used to install GL/gl.h & co. only if libGL.so was installed.

After this commit
https://cgit.freedesktop.org/mesa/mesa/commit/?h=13.0&id=a89faa2022fd995af2019c886b152b49a01f9392

this is no longer the case leading to misdetection of opengl support by
vlc if mesa3d was built before:

http://git.videolan.org/?p=vlc.git;a=blob;f=configure.ac;h=c8900d216540816224e872bef46fbcc93d83ca41;hb=HEAD#l3010

This patch partly re-introduces a mechanism to clean unwanted files
installed by mesa3d which was removed when bumping to 12.0.0:
https://git.buildroot.net/buildroot/commit/package/mesa3d?id=ff3e522e748925f6be5732a6e5df6553468ca0a1

Fixes
http://autobuild.buildroot.net/results/13e/13e7c0f5fc6ae8ba5e62b2597d188496c64090fd/

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

pkg-cmake: Change -DNEBUG to -DNDEBUG

Signed-off-by: Stefan Nickl <Stefan.Nickl@gmail.com>
Acked-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

jamvm: fix build with uClibc on i386/x86_64

Commit 0b6b67f34d6d4aeb340bbca579dbf85363c3f4ea ("jamvm: add patch to
fix musl build") introduced a patch to fix the jamvm build with the
musl C library. While the commit log pretends that the build was still
working with uClibc, it is not correct: it no longer builds fine with
uClibc on i386/x86_64, because the Buildroot default configuration for
uClibc doesn't enable <fenv.h> support.

Therefore this commit adapts the patch to use <fenv.h> if available
(which is the case with musl), and otherwise fall back to
<fpu_control.h>, which is available in uClibc.

Thanks to Waldemar for the investigation.

Fixes:

  http://autobuild.buildroot.net/results/325a50d15e1836b31df4e84ba83b296abfb73041/

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

package/ser2net: disable threads for toolchains without threads support

ser2net configure script provide an option to disable threads support,
use it.

Fixes:
http://autobuild.buildroot.net/results/6bc/6bc0f99917c2069a1614f4cfcaa98af1b21762e4

Signed-off-by: Romain Naour <romain.naour@gmail.com>
[Thomas: use --{with,without}-pthreads instead of --with-pthreads={yes,no}.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

privoxy: needs thread support

Even though privoxy has some conditional code to handle configurations
without threads, it is not properly handled everywhere, leading to a
build failure. Since this package is unlikely to be used in
non-threaded environment, just disable it for non-threaded
configurations.

Fixes:

  http://autobuild.buildroot.net/results/fae7a41d110ed6c4d62c77873229d069d8b65986

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

libnss: add upstream patch to fix uninitialized variable warning

Fixes:

  http://autobuild.buildroot.net/results/47e611036be12844ebca093ddd8fe95c326c6056/

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

binutils: add patch to 2.27 to fix arm issue with file having no data

This commit adds a patch to binutils 2.27, which fixes the build of
kvmtool on ARM. The fix is taken from the upstream mailing list at
https://lists.gnu.org/archive/html/bug-binutils/2016-08/msg00165.html.

Fixes:

  http://autobuild.buildroot.net/results/bae83d495b46efeec45c8008949c811b6241adaa

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

kvm-unit-tests: add patch to remove -Werror

Fixes:

  http://autobuild.buildroot.net/results/52c02c68be6c0b3e907f7ed645bd2d18476ec4f3/

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

kvm-unit-tests: really use KVM_UNIT_TESTS_ARCH and fix value on x86-64

The kvm-unit-tests package defines a KVM_UNIT_TESTS_ARCH variable, but
never used it. Due to this, on PowerPC64, we were passing an incorrect
value, causing a build failure.

In addition, once the KVM_UNIT_TESTS_ARCH is used, it appeared that
its value for x86_64 was incorrect, due a typo: it was x86_84 instead
of x86_64.

Fixes:

  http://autobuild.buildroot.net/results/16836c02f4d1ab5df6b2e21a4fcdbbf4a7542f22/

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

binutils: add upstream patch to fix open microblaze issue

This reenables the use of 2.27 for microblaze.

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

package/pseudo: update version

Brings in a fix about externded attributes (xattrs).

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

configs/kb9202_defconfig: remove defconfig

This configuration for the KwikByte KB9202 was added back in July 2008
(commit e27fef51834042a1b108e04927a2a91742a54327), and was never
updated since then. It still uses a 2.6.38 kernel, with the out of
tree patches from the former AT91 kernel maintainer. The board has
been discontinued according to its page at
http://www.kwikbyte.com/KB9202.html.

This defconfig no longer builds since the switch to gcc 5.x as the
default. Therefore, it's time to get rid of this defconfig.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

makedevs: make device node creation idempotent

We use makedevs to create device nodes in the target rootfs. However,
this can be called several times, e.g. when building several filesystem
images or when rebuilding. When makedevs is called the second time, the
device node already exists so mknod() errors out.

This wasn't noticed before because fakeroot's mknod() wrapper
(incorrectly) does _not_ error out when the file exists already. Now
we switched from fakeroot to pseudo, the problem becomes apparent.

Before creating the device node, check if it already exists and if so,
if it has the correct device type and number. Change of mode and
ownership is still done.

This approach was preferred over removing the target files before
creating them, which would be simpler. However, when e.g. a file exists
as a normal file and makedevs specifies it as a device node, that
really is an error so we should detect it.

The other types don't have to be changed. The 'd' (directory) type is
already OK because it already only creates directories if they don't
exist yet. The 'f' (file mode) and 'r' (recursive) types only operate
on files and directories that exist already.

Patch also sent upstream to busybox.

Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Reported-by: Fabio Estevam <festevam@gmail.com>
Tested-by: Fabio Estevam <festevam@gmail.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

makedevs: resync device creation with upstream busybox

In upstream busbyox, the code to create devices has been simplified:
the code for a single and for multiple devices is no longer duplicated.

Since we are going to change the device creation code next, it's
convenient to have only one copy to modify.

There are two behavioural changes with this, but they were introduced
silently together with other commits in upstream busybox.

- When mknod() fails, the chmod was still done. This is pointless so it
  is no longer done now.

- There was a check for mode != -1; however, a mode of -1 would not
  have worked anyway because all bits would be set for mknod(), which
  would fail. So this check is redundant.

Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Tested-by: Fabio Estevam <festevam@gmail.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

libxml2: add upstream security fixes

Fixes: CVE-2016-5131, CVE-2016-4658.
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

czmq: fix build with glibc 2.24

czmq uses readdir_r(), which has been marked as deprecated in glibc
2.24, and czmq uses -Werror which causes the warning to be turned into
an error that aborts the build.

This commit fixes that by removing -Werror. Long-term, we might want
to upgrade czmq to a newer version, since they've removed the use of
readdir_r() in commit 2594d406d8ec6f54e54d7570d7febba10a6906b2. But
getting rid of -Werror is anyway always a good idea in Buildroot.

Fixes:

  http://autobuild.buildroot.net/results/f6e4b8ed8571fdcf30eecae121dd11dfdef50799

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

packages/pseudo: needs attr

This fixes the build on machines that do not have attr development files
installed.

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

pseudo: fix build for python3 based distros

Pseudo is not python3 friendly. It causes build failure on distros using
python3 as default python interpretor.

./maketables enums/*.in
  File "./makewrappers", line 327
    return """/* This function is not called if pseudo is configured --enable-force-async */
                                                                                           ^
TabError: inconsistent use of tabs and spaces in indentation
  File "./maketables", line 76
    print "Flags: set for %s" % self.name
                            ^
SyntaxError: Missing parentheses in call to 'print'
make[2]: *** [Makefile:150: wrappers] Error 1
make[2]: *** Waiting for unfinished jobs....
make[2]: *** [Makefile:147: tables] Error 1

Those patches make pseudo works with python2 and python3.

Signed-off-by: Gaël PORTAY <gael.portay@savoirfairelinux.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

tinyalsa: avoid install in /usr/local/

Set the PREFIX to avoid installing files in [...]/usr/local/.

Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

lttng-babeltrace: fix static build again

I accidentally removed the static build patch when bumping to v1.4.0,
thinking it was merged upstream as of this version, but it's not the
case (it's merged in master), so let's keep the patch file.

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

configs/nanopi-neo: use a more conventional kernel

It turns out that linux-4.9-rc3 has all that is needed for the NanoPi
NEO, so use that rather than the franken-kernel I assembled.

Reported-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/mesa3d: Fix endianess detection with musl-based toolchains

Fixes
http://autobuild.buildroot.net/results/e27/e27a9a95f72dba3076549beb2a2ccfdbea2fcfee/
http://autobuild.buildroot.net/results/462/4629ee4813aa36de884bba2576612f909c007264/

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

get-developers: fix parentheses for print

Python3 complains about missing parentheses.

$ ./support/scripts/get-developers
  File "./support/scripts/get-developers", line 45
    print f
          ^
SyntaxError: Missing parentheses in call to 'print'

Signed-off-by: Gaël PORTAY <gael.portay@savoirfairelinux.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>