git.libre-soc.org Git - buildroot.git/log

bctoolbox: fix mbedtls/polarssl dependency

Currently we force selection on mbedtls for bctoolbox. If, for whatever
reason, mbedTLS is not found, bctoolbox searches for PolarSSL and, if
not present, fails with:

```
Could NOT find PolarSSL (missing: POLARSSL_INCLUDE_DIRS
HAVE_POLARSSL_SSL_H)
```

This happens because bctoolbox enables support for mbedTLS and PolarSSL
by default, whereas, if both are enabled, mbedTLS is preferred over
PolarSSL.

However, crypto support is not mandatory and support for both libraries
can be turned off. Therefore, optionally depend on mbedTLS or PolarSSL
and enable/disable the library support explicitly.

This fixes the issue of searching for PolarSSL altough the dependency is
not declared in the bctoolbox package.

Note, that this does not fix the issue why mbedtls is not found although
it is enabled. This issue is address in this thread [1].

Fixes:
http://autobuild.buildroot.net/results/cfe/cfeb2f542598e5d450b332fb51a6d79bae24158c/
.. and more.

[1] http://lists.busybox.net/pipermail/buildroot/2017-February/183055.html

Signed-off-by: Jörg Krause <joerg.krause@embedded.rocks>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

bctoolbox: add patch to fix static linking

Static linking with bctoolbox fails when using pkg-config as the
generated bctoolbox.pc file only consists of an '-L' string without any
search path or libraries flags:

```
libtool: link: need path for `-L' option
```

That's because of an typo in mbedtls_library_path. However,
mbedtls_library_path contains a string of the mbedtls libraries
concatenated by an ';' which cannot be parsed by pkg-config.

Therefore, use MBEDTLS_LIBRARY instead of MBEDTLS_LIBRARIES to get
the library path.

Furthermore, add the three mbedtls libraries mbedtls, mbedcrypto, and
mbedx509 to LIBS_PRIVATE so these libraries are added to the
Libs.private field of bctoolbox.pc.

Fixes:
http://autobuild.buildroot.net/results/37d5625df4be11ccdc063871e9f6e13d5f59fb52
http://autobuild.buildroot.net/results/1999c841fae41f860f00747a362327cb2857e687
.. and many more.

Upstream status: Pending
https://github.com/BelledonneCommunications/bctoolbox/pull/4

Signed-off-by: Jörg Krause <joerg.krause@embedded.rocks>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

Revert "bctoolbox: fix typos and logic error"

This reverts commit 5c4d3560b94c9cb8f1594c9200a632b89ddf7aa0.

Although the patch fixes the build error it does not set the flags for
Libs.private properly. A follow-up patch will provide a better solution
for this issue.

Signed-off-by: Jörg Krause <joerg.krause@embedded.rocks>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

packages: improve license type lists

Make license type lists more uniform:

* put content license applies to in parenthesis; ex: "GPLv2+ (programs)"
* use commas to separate types listed without conjuction; ex: "GPLv2, LGPLv2"

No attempt was made to validate the claimed licenses. This is just a tweak
to increase uniformity of the _LICENSE variables.

Signed-off-by: Danomi Manchego <danomimanchego123@gmail.com>
Reviewed-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
[Thomas: replace semi-colons by commas in LIBURCU_LICENSE.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

qt5base: use system/buildroot provided freetype

Select (and use) builroot provided freetype in case qt5base gui is
selected (gui without freetype/font support makes not much sense
and gives compile errors).

Qt 5.6 needs some support to find the include headers, 5.8 is
clever enough inherently.

Fixes [1] (by avoiding the build-in freetype library which depends
on activated libpng support):

Project ERROR: Library 'libpng' is not defined.
make[3]: *** [sub-3rdparty-freetype-make_first] Error 3

[1] http://autobuild.buildroot.net/results/219162ba616289f799a5053c4dcc6b8574284283

Signed-off-by: Peter Seiderer <ps.report@gmx.net>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

qt5base: move linuxfb configure option comment

Move the two comment lines to where they apply.

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

util-linux: build schedutils without enabling basic binaries

Pull an upstream fix to allow building schedutils (ionice, taskset,
chrt) when util-linux is configured with

$ configure --disable-all-programs --enable-schedutils

Fixes:
https://bugs.busybox.net/show_bug.cgi?id=9656

Reported-by: Aaron Holtzman <aholtzma@gmail.com>
Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

directfb: add gfx-state initialisation

The patch 0006-fix-client-gfx_state-initialisation.patch
add the initialisation for the variable gfx-state.

This patch was provided by the community and it is already
integrated in OpenEmbedded.

Signed-off-by: Philippe Reynes <philippe.reynes@sagemcom.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

uclibc: add fix for ARM specific static linking issues

Fixes:
http://autobuild.buildroot.net/results/fb14e37aaf453874d1c33e5ed73b9c751ace5ae3
http://autobuild.buildroot.net/results/774a8dd489760bc213ac7cf3e8040cee6a4e2437

And hopefully all other static ARM issues with uClibc-ng.

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

uclibc: define struct ucontext for or1k

Add missing define for struct ucontext for OpenRISC.

Fixes:
http://autobuild.buildroot.net/results/d8715a2b3783d877edf91bb14a80d36eac749b36

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

qt5: disallow 5.8 version for ARMv4

Fixes:
http://autobuild.buildroot.net/results/197/197ab8a12bf1d780a986e5d447d1e0f5ee070405/

qt5.8 no longer supports ARMv4. From src/corelib/global/qprocessordetection.h:

  if Q_PROCESSOR_ARM >= 6
    define Q_PROCESSOR_ARM_V6
  endif
  if Q_PROCESSOR_ARM >= 5
    define Q_PROCESSOR_ARM_V5
  else
    error "ARM architecture too old"

[Peter: also hide comment as pointed out by Baruch]
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

erlang: set ERL_TOP to avoid using user's version

Explicitly set the ERL_TOP environment variable to override values set
on a user's system. This prevents Erlang's configure scripts from
using source files outside of the ones that Buildroot downloaded.

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

wavpack: properly wrap help text

Wrap help text to fit 72 columns.

Signed-off-by: Jörg Krause <joerg.krause@embedded.rocks>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

wavpack: fetch upstream patch to remove wchar dependency

The recent bump to version 5.1.0 added a dependency on wchar:

```
import_id3.c:37:34: error: unknown type name 'wchar_t'
```

This issue has been reported upstream [1] and a fix was commited by
upstream to remove the wchar dependency [2].

Fixes:
http://autobuild.buildroot.net/results/9a6/9a693f5b798571917f36cfb7661e2f2638aac550/
http://autobuild.buildroot.net/results/44c/44c8227043045baf4f043da44b8129e43dfff687/
http://autobuild.buildroot.net/results/a80/a80221dcc0860046ebdf0bbf454e056b1e20df83/
.. and more.

The build was successfully tested with the 'br-arm-basic-2017.02-rc1-2-g133c5ac'
toolchain.

[1] https://github.com/dbry/WavPack/issues/19
[2] https://github.com/dbry/WavPack/commit/876fc3f3907e871d0938ac6c8c5252f5f31abd1f

Signed-off-by: Jörg Krause <joerg.krause@embedded.rocks>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/glmark2: needs gcc >= 4.9 for C++14

Fixes:
http://autobuild.buildroot.org/results/511/51162285d197deabb63727c19cdf5512b0e68f29/

[Peter: also add libstdcpp dependency]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Bernd Kuhls <bernd.kuhls@t-online.de>
Cc: Spenser Gilliland <spenser@gillilanding.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/mosquito: needs sync_4

Fixes:
    http://autobuild.buildroot.org/results/2bc/2bc84ba2d1167018e2d48e5183ead22b6425dcf5/
    http://autobuild.buildroot.org/results/445/445f377ae70397b5f675f541977900e8986b79a4/
    http://autobuild.buildroot.org/results/57e/57e6984427f8c5d906a93884cc461b8f93cf5ce0/
    ...

[Peter: also add dependency to mosquitto comment]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/mpv: fixup fix build on powerpc64 w/ altivec

The initial fix for this issue, re-including stdbool.h, is not
working.

Instead, fix using the method used for other packages with a similar
problem: add -std=gnu99, which prevents bool from being redefined and
removes the need for any source code change.

Fixes: 64904f0f6d691cfea5ae1c8f58e2d61aac0684c2
Signed-off-by: Sam Bobroff <sam.bobroff@au1.ibm.com>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

kmod: workaround for microblaze ld bug

Add a workaround for a microblaze specific ld bug to avoid
segfault.

Fixes:
http://autobuild.buildroot.net/results/aef1b915ffd3e5678af8682727c22c5618edfd89

[Peter: force --gc-sections configure check to false instead of patching
configure]
Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libnss: workaround for microblaze ld bug

Add a workaround for a microblaze specific ld bug to avoid
segfault.

Fixes:
http://autobuild.buildroot.net/results/638a65453879777a0d5bdb29231034cd261b41c0

[Peter: rename to LIBNSS_DROP_GC_SECTIONS for clarity]
Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

DEVELOPERS: adopt freerdp

Julian only ever sent one patch to freerdp, the one adding it, in 2012.

I'm somewhat interested in this package, so I'll adopt it.

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

gstreamer1: drop unused configure time unaligned access handling logic

gstreamer has moved to build time unaligned access support detection using
ifdefs in gstconfig.h since 1.9.2:

commit 6ef601367e0f5adb7a8b02cec94dec04fccf701a
Author: Nirbheek Chauhan <nirbheek@centricular.com>
Date:   Tue Jun 21 18:59:49 2016 +0530

    gstconfig.h: Detect unaligned access support at compile-time

    This makes gstconfig.h completely arch-independent. Should cover all
    compilers that gstreamer is known to build on, and all architectures
    that I could find information on. People are encouraged to file bugs if
    their platform/arch is missing.

So ac_cv_unaligned_access is no longer used and our logic can be removed.

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

gstreamer1: fix unaligned detection for or1k

Fixes:
http://autobuild.buildroot.net/results/717/717d78ce0935749f477bdf3133b6f20057a28c01/

Add patch to fix unaligned detection for or1k. Patch submitted upstream:

https://bugzilla.gnome.org/show_bug.cgi?id=778866

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

gstreamer: fix unaligned detection for or1k

Similar to how we do it for the other niche/new architectures.

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

samba4: requires NPTL

With OpenRISC there is a uClibc-ng based toolchain for an
architecture with MMU support, but with only Linuxthreads
and no NPTL. Samba4 uses pthread_mutexattr_setrobust which
is not implemented in uClibc-ng Linuxthreads.

Fixes:
http://autobuild.buildroot.net/results/ec8f078046c92ae23dc973c88ac0fef136ad1535

Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/mpd: use proper BR2_PACKAGE_MPD_VERSION_0_19 symbol

Commit 9244ad854 [1] adds a hidden choice between version 0.19 and 0.20. The
Makefile now uses a wrong symbol introduced by this commit. Fix by:
s/BR2_MPD_VERSION_0_19/BR2_PACKAGE_MPD_VERSION_0_19.

[1] https://git.buildroot.net/buildroot/commit/?id=9244ad854d30e8c942460646ef6f6194ee44942d

Signed-off-by: Jörg Krause <joerg.krause@embedded.rocks>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

dbus: security bump to version 1.10.16

>From http://www.openwall.com/lists/oss-security/2017/02/16/4

The latest dbus release 1.10.16 fixes two symlink attacks in
non-production-suitable configurations. I am treating these as bugs
rather than practical vulnerabilities, and very much hope neither of
these is going to affect any real users, but I'm reporting them to
oss-security in case there's an attack vector that I've missed.

No CVEs assigned so far.

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

trousers: disable on ARC

This package has been failing to build on ARC due to a toolchain bug
since December 2015, causing a total of 168 build failures. Therefore,
let's disable this package for now, until the toolchain issue gets
resolved.

Fixes^Wavoids:

http://autobuild.buildroot.net/results/e65d1d7c5a4dd4616167cfe2aa938be6af139602/

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

libraw: fix build with old glibc versions

Prior to glibc 2.18, definitions like SIZE_MAX or INT_FAST32_MAX from
<stdint.h> were only made available for C code, or in C++ if
__STDC_LIMIT_MACROS was defined.

The code from jasper uses such definitions, without defining
__STDC_LIMIT_MACROS. Unfortunately, defining __STDC_LIMIT_MACROS in
the jasper headers doesn't work, since <stdint.h> has already been
included before, at a point where __STDC_LIMIT_MACROS was not defined.

So to solve this problem, we simply pass -D__STDC_LIMIT_MACROS in
CXXFLAGS when building libraw with jasper support.

Fixes:

http://autobuild.buildroot.net/results/908aef6c82d56060933713df217b6b2ba21a01b0/

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

tcping: fix upstream location

The upstream site is no longer available from http://www.linuxco.de,
but only from http://linuxco.de.

Fixes:

http://autobuild.buildroot.net/results/03caf8403c31e2424c338de4df636dd900408b14/

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

postgresql: disable spinlocks on openrisc

Just like all non-mainstream architectures, postgresql doesn't provide
its internal spinlock implementation on OpenRISC, so disable it.

Fixes:

http://autobuild.buildroot.net/results/e9de750e34f10c4a2403696936260ad1547d0d15/

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

ocrad: use generic extract command

The generic code now knows how to extract .tar.lz archives. Remove the local
extract code

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

ddrescue: use generic extract command

The generic code now knows how to extract .tar.lz archives. Remove the local
extract code

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

ed: use generic extract command

The generic code now knows how to extract .tar.lz archives. Remove the local
extract code.

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

package: add generic support for lz archives

This commit teaches the generic package handling code how to extract .tar.lz
archives. When lzip is not installed on the host, host-lzip gets built
automatically.

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

package: refactor listing of extractor dependencies

Don't special case $(XZCAT) when constructing DL_TOOLS_DEPENDENCIES. The next
commit will introduce another extractor that automatically builds when not
installed. Introduce EXTRACTOR_DEPENDENCY_PRECHECKED_EXTENSIONS that lists
archive extensions for which the extractor is already checked in
support/dependencies/check-host-foo.mk. Use this in the newly introduced
extractor-dependency to populate DL_TOOLS_DEPENDENCIES.

Cc: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Reviewed-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
[Thomas: add missing space after "firstword", as noticed by Thomas DS.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/mpd: add choice between version 0.19 and 0.20

MPD version 0.20 being affected by GCC PR libstdc++/64735 means no mpd package
available in Buildroot for the architectures NIOSII, ARMv4, ARMv5 and SPARCv8
until GCC 7 is released.

As the next Buildroot release is in 2017.02, which is before GCC 7 is expected
to be finished we add hidden config symbols to select the version branch to use.

Note, that we bumped the version from 0.19 to 0.20 quite recently within the
current Buildroot release cycle. This way we can keep the support for mpd on
the affected architectures, like ARMv5, until GCC 7 is out.

Note, that for the version 0.19 MPD requires libglib2.

Signed-off-by: Jörg Krause <joerg.krause@embedded.rocks>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

cannelloni: fix build failure caused by missing include

Add a patch, that explicitly includes cstdio header.

Fixes:
http://autobuild.buildroot.net/results/82d/82d1dca3aa7ee9189baae985db59464d95696ed7/

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

board/udoo/neo: specify an offset for the rootfs partition

When manually writing SPL and u-boot.img to the SD card the rootfs
gets corrupted.

Fix this issue by passing a 1MB offset for the rootfs.

Reported-by: Breno Lima <breno.lima@nxp.com>
Signed-off-by: Fabio Estevam <fabio.estevam@nxp.com>
Tested-by: Breno Lima <breno.lima@nxp.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

linux-headers: bump 4.{4, 9}.x series

Signed-off-by: Fabio Estevam <fabio.estevam@nxp.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

linux: bump default to version 4.9.10

Signed-off-by: Fabio Estevam <fabio.estevam@nxp.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

mesa3d-headers: remove installation of OpenVG headers

OpenVG support has been completely removed from mesa3d in commit
3acd7a34ab05b87521b74f626ec637e7fdcc6595, merged in mesa-10.6.0.

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>

mesa3d-headers: fix installation of dri.pc

/usr/lib/pkg-config/ is not the correct installation path for .pc files,
it should be in /usr/lib/pkgconfig/

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>

package/x11r7/xserver_xorg-server: Add patch to fix v1.19.1 with SIGIO

The commit 6a5a4e60373c from xserver v1.19.1 removes SIGIO support for
input [1].

The current patch coming from Open Embedded [2] removes a check of
'useSIGIO'. This check was producing a segfault error while using
X11 and Vivante driver [3]. Thanks to Gary Bisson to point out this
patch!

[1] https://cgit.freedesktop.org/xorg/xserver/commit/?id=6a5a4e60373c1386b311b2a8bb666c32d68a9d99
[2] https://patchwork.openembedded.org/patch/136696/
[3] http://lists.busybox.net/pipermail/buildroot/2017-February/183994.html

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

linux-headers: bump 3.10.x series

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

redis: bump to version 3.2.8

It fixes a regression in the 3.2.7 security release that can cause server
deadlocks.

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

musl: no SSP on i386 and PowerPC

Due to what appears a bug in gcc according to the musl developers, but
that the gcc developers don't want to fix, SSP support currently don't
work on i386 and PowerPC with musl.

Additional details can be found at:

  http://www.openwall.com/lists/musl/2016/12/04/2

OpenWRT and Alpine Linux both have musl and gcc patches to work around
the issue, but in the context of Buildroot, we at this point don't care
enough about SSP support specifically with musl on those architectures
to carry additional patches.

Currently, having SSP enabled with musl/i386 causes a number of build
failures in the autobuilders: cups, ipmiutil, openssh, ruby, stunnel,
sudo and mosh at least all fail to build because of this.

So we simply disable SSP support in the toolchain when musl is used on
i386 and PowerPC. The PowerPC case is not tested in the autobuilders,
but has been reproduced locally and is also fixed by this patch.

Fixes:

  mosh
  http://autobuild.buildroot.net/results/60aa12f1aed08e3b7a98f9ce7091bee3a44d692c/

  ipmi-util
  http://autobuild.buildroot.net/results/fb9a071b8739527f424cfe2886ec480f438f70ab/

  cups
  http://autobuild.buildroot.net/results/486dea944d6ecba5c4e6e8ac664261c1909f4b4c/

  openssh
  http://autobuild.buildroot.net/results/742a8bf4726de6e9ba6926e3fb6019a434454e48/

  sudo
  http://autobuild.buildroot.net/results/682531f368c4e982cafe9e625dd41f6d8c7f93f9/

  ruby
  http://autobuild.buildroot.net/results/dac660f96c7f85e933a6b82cf61edd429eeae9aa/

  stunnel
  http://autobuild.buildroot.net/results/cee52505f1ac2da2f5ba86c9ebfd1f5cd9e301be/

Thanks to Yann E. Morin for suggesting to simply disable SSP support
rather than trying to fix it.

[Peter: add comment explaining why]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/libgpiod: bump version to 0.1.3

Release 0.1.2 claimed to fix an issue with requesting both types of
line events, but actually didn't. This time it's really fixed.

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

ddrescue: fix uclibc compile

Fixes compile failure with uclibc [1] because of missing stdio.h include:

  In file included from io.cc:28:0:
  block.h:219:22: error: ‘FILE’ has not been declared
     int write_mapfile( FILE * f = 0, const bool timestamp = false,

[1] http://autobuild.buildroot.net/results/4ac0754f1cc5ea934d6437e89d1f4906fb3fd0a8

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

libv4l/ir-ctl: fix musl compile

- add copy of TEMP_FAILURE_RETRY macro
- use strdup instead of strdupa

Fixes [1]:

ir-ctl.c:(.text+0xb06): undefined reference to `strndupa'
ir-ctl.c:(.text+0xe01): undefined reference to `TEMP_FAILURE_RETRY'

[1] http://autobuild.buildroot.net/results/b8b96c7bbf2147dacac62485cbfdbcfd758271a5

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

cups: add host-pkgconf as a dependency

cups configure script checks for a pkg-config program. If host-pkgconf
has not been built prior to cups, it uses the system pkg-config, which
might potentially lead to bogus results:

checking for i586-buildroot-linux-musl-pkg-config... no
checking for pkg-config... /usr/bin/pkg-config

With this patch applied:

checking for i586-buildroot-linux-musl-pkg-config... no
checking for pkg-config... /home/thomas/projets/buildroot/output/host/usr/bin/pkg-config

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

cbootimage: disable on musl

cbootimage uses non-standard u_intXX_t types instead of the C99 types
uintXX_t types, which prevents it from building with the musl C library.

A patch to fix this has been submitted upstream, but since there are
many many occurences of such types, the patch is a bit large to carry in
Buildroot, especially compared to the fact that having this package
building against musl is probably not a top priority.

So for now, simply disable this package on musl, until upstream fixes
the issue and releases a new version.

Fixes:

http://autobuild.buildroot.net/results/61bdfb7e0ff9628190d9eb86e40c4c90e768b8e2

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

package/hiredis: fix makefile code

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

package/ccache: do not force colored diagnostics

When GCC_COLORS is set, ccache passes '-fdiagnostics-color' to GCC but
this flag requires GCC v4.9 or later. Older versions complain about the
unrecognized command line option.

Using GCC_COLORS in the context of Buildroot is seldom useful, so we
just unexport GCC_COLORS altogether.

Reported-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

kmsxx: fix build error for m68k/coldfire

Fixes:
http://autobuild.buildroot.org/results/fc8/fc820e0b558537f771588b0e1291a8332427ff8d/

Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libasplib: add url to gcc bug report

Add comment to bug report.

Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

lcdapi: fix build error for m68k/coldfire

Fixes:
http://autobuild.buildroot.net/results/d47/d47fa41aa860d82471b83ac90967d3a3dacd8611/

Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

openswan: fix autobuild failures with uClibc-ng

Add patch to fix following autobuild failures:

Fixes:
http://autobuild.buildroot.net/results/31cf680c48bec8ba4450a6fe2156cb441c4dd660
http://autobuild.buildroot.net/results/95220f231810cb12ba97a6b44af79f84af706c82

Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

ntfs-3g: add security fix for CVE-2017-0358

Jann Horn, Project Zero (Google) discovered that ntfs-3g, a read-write
NTFS driver for FUSE does not not scrub the environment before
executing modprobe to load the fuse module. This influence the behavior
of modprobe (MODPROBE_OPTIONS environment variable, --config and
--dirname options) potentially allowing for local root privilege
escalation if ntfs-3g is installed setuid.

Notice that Buildroot does NOT install netfs-3g setuid root, but custom
permission tables might be used, causing it to vulnerable to the above.

ntfs-3g does not seem to have a publicly available version control system
and no new releases have been made, so instead grab the patch from Debian.

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

vim: security bump to version 8.0.0329

Fixes:

- CVE-2016-1248: vim before patch 8.0.0056 does not properly validate values
  for the 'filetype', 'syntax' and 'keymap' options, which may result in the
  execution of arbitrary code if a file with a specially crafted modeline is
  opened.

- CVE-2017-5953: vim before patch 8.0.0322 does not properly validate values
  for tree length when handling a spell file, which may result in an integer
  overflow at a memory allocation site and a resultant buffer overflow.

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

bctoolbox: fix typos and logic error

To get the directory path from the library name you need to
use a single filename. Fix typo in variable name.

This fixes the ortp autobuild errors which uses the broken pkgconfig file
from bctoolbox.

Fixes:
http://autobuild.buildroot.net/results/37d5625df4be11ccdc063871e9f6e13d5f59fb52
http://autobuild.buildroot.net/results/1999c841fae41f860f00747a362327cb2857e687

Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

glmark2: bump version to fix build

The package otherwise fails to build with a recent toolchain with GCC6
(tested with Linaro ARM 2016.11).

It used to fail at sqrt check during package configuration:
Checking for function sqrt : not found
The configuration failed

Bumping version to latest HEAD fixes the issue as explained in the
following discussion:
https://github.com/glmark2/glmark2/issues/15

Fixes:

http://autobuild.buildroot.net/results/8dc9400505b9087ce290981d95486598df0beb56/

Signed-off-by: Gary Bisson <gary.bisson@boundarydevices.com>
[Thomas: add reference to autobuilder failure.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/classpath: Don't depend on jamvm

GNU classpath can exist without jamvm, so don't depend on it.

Signed-off-by: Marcus Hoffmann <m.hoffmann@cartelsol.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

wget: bump to 1.19.1 to fix build issue

Due to the patches we have on wget 1.19, we need to
autoreconf. Unfortunately, when the autoreconfiguration process occurs
with host-gettext already built and installed, the build of wget fails
with a fairly weird error:

In file included from str-two-way.h:44:0,
from c-strcasestr.c:37:
./stdint.h:89:5: error: #if with no expression
#if

As explained in http://git.net/ml/bug-gnulib-gnu/2017-01/msg00067.html
and the links pointed by this page, this is due to an incompatibility
between the newer version of gnulib used in wget, and an older .m4 file
in gettext.

In the context of Buildroot, the easiest way to avoid the issue is to
not autoreconf wget. The wget project has conveniently released a 1.19.1
release, which contains our two patches, plus just one small feature
addition. It is therefore reasonable to apply this as a solution to this
build issue.

Fixes:

http://autobuild.buildroot.net/results/b62ac6fd5ce36453935c309e112262467cf0e3bf/

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

xfsprogs: needs thread support

Fixes:

http://autobuild.buildroot.net/results/57975f401f39b673eeec304b4738cfb41a19ece7/

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

docs/website/news.html: fix link to 2017.02-rc1 tarball

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

riemann-c-client: disable parallel make

Fixes:
http://autobuild.buildroot.net/results/2f8/2f85a2d19232365f7e5fddde6092af9dd74d4518/
http://autobuild.buildroot.net/results/80d/80d83c650c668ee1e87c288bd7a0ce63eab95631/

The build system doesn't specify any dependencies between the generated
header files and the source files including them, causing a race condition:

  GEN      lib/riemann/proto/riemann.pb-c.c
  GEN      lib/riemann/proto/riemann.pb-c.h
  CC       src/src_riemann_client-riemann-client.o
In file included from ./lib/riemann/riemann-client.h:23:0,
                 from src/riemann-client.c:18:
./lib/riemann/attribute.h:21:40: fatal error: riemann/proto/riemann.pb-c.h: No such file or directory

Work around it by forcing non-parallel make.

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

bind: security bump to version 9.11.0-P3

Fixes CVE-2017-3135: Combination of DNS64 and RPZ Can Lead to Crash:

https://kb.isc.org/article/AA-01453

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

support/test-pkg: report number and types of failures

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Luca Ceresoli <luca@lucaceresoli.net>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

support/test-pkg: be less verbose

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

support/test-pkg: simplify the config check

Use comm(1) to check that all our config options are properly set in the
resulting configuration, rather than our canned and fragile code.

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

support/test-pkg: check the config snippet exists

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

support/test-pkg: fix code style

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

wiringpi: fix static linking

Update patch 0001-Adjust-for-buildroot-build.patch to
use real static linking (command line flag '-static' for
target gpio-static).

Fixes [1]:

  [Link]
  .../output/host/usr/lib/gcc/powerpc-buildroot-linux-uclibc/5.4.0/libgcc.a(unwind-dw2-fde-dip.o): In function `_Unwind_Find_FDE':
  ...output/build/host-gcc-final-5.4.0/build/powerpc-buildroot-linux-uclibc/libgcc/../../../libgcc/unwind-dw2-fde-dip.c:465: undefined reference to `dl_iterate_phdr'
  collect2: error: ld returned 1 exit status
  make[1]: *** [gpio-static] Error 1
  make[1]: Leaving directory `.../output/build/wiringpi-2.32/gpio'

[1] http://autobuild.buildroot.net/results/b905617dda0e120d04c54daf7cae2bad236ac767

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

package/gdb: backport upstream fix for simulator build issue with bfin

This revert the commit [1] which break the simulator for aarch64.
Backport the upstream fix instead (from 7.12).

[1] f71ad71f24fbeca1a81cc9efe79016a7db6f1320

Tested with gcc 4.4.5 (debian squeeze chroot).

Fixes:
http://autobuild.buildroot.net/results/cfd/cfdc9117fef7ecdf5cc5fc907a6fe8701a2c174a

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Cc: Waldemar Brodkorb <wbx@openadk.org>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

DEVELOPERS: update my entry list

Interested in all qt5 stuff.

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

docs/website: update release announcement with mailing list link

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

CHANGES: final updates before 2017.02-rc1

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

Prepare for 2017.02-rc1

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

CHANGES: prepare for 2017.02-rc1

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

poppler: qt5 support needs c++11 (since Qt5.7)

Signed-off-by: Andreas Naumann <anaumann@ultratronik.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

poppler: Add option for Qt5 support

Add an option to build the libpoppler-qt5 library.

Signed-off-by: Andreas Naumann <anaumann@ultratronik.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

qt5: Rename version variable to include _PACKAGE_

Signed-off-by: Andreas Naumann <anaumann@ultratronik.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

motion: disable autotools host optimization

The motion autotools script helpfully detects the host
Microprocessors intr set and optimizes for it.
Disabling this feature causes autotools to use the
target parameters passed by BR instead.

Signed-off-by: Ray Kinsella <ray.kinsella@intel.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

wget: add upstream patch to fix build failure

Fixes:
http://autobuild.buildroot.net/results/c86/c8657563e63e1012a3ae3c0c47663a951e280022

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

util-linux: enable AUTORECONF

Commit 55a9737895cb ("util-linux: link scriptreplay with libm (for isnan)")
added two patches that touch configure.ac and Makemodule.am. But forgot
to enable AUTORECONF.

When AUTORECONF is disabled and configure.ac is patched, it looks like
make will detect change in timestamps and trigger reconfig. But it
later fails because of missing dependencies. To fix this, explicitly
enable AUTORECONF.

Fixes:
http://autobuild.buildroot.net/results/544/544e8da290d40424ea3d1bffad7e0b8a566de495

Fixes: 55a9737895cb ("util-linux: link scriptreplay with libm (for isnan)")
Signed-off-by: Rahul Bedarkar <rahul.bedarkar@imgtec.com>
Cc: Carlos Santos <casantos@datacom.ind.br>
Acked-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

ddrescue: needs c++ compiler

Fixes [1].

[1] http://autobuild.buildroot.net/results/0af569ad06efa36e8e1ae45acd1facfd98b86f20

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

mkpimage: bump to latest barebox version

Adds support for header v1 (used on Arria 10 SoCs)

Signed-off-by: Lionel Flandrin <lionel@svkt.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

lxc: bump to version 2.0.7

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

support/test-pkg: add option to limit the number of tests

Sometimes, it interesting to have a global overview of whether the
package builds at all or not, rather than test on all toolchains.

Add an option that allows testing on a limited set of randomly choosen
toolchains.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Luca Ceresoli <luca@lucaceresoli.net>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

support/test-pkg: store lines missing from resulting configuraiton

When a build is skipped, store the lines from the config snippet, that
are missing in the resulting configuration, in a file in the build
directory, for the user to inspect.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Luca Ceresoli <luca@lucaceresoli.net>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Acked-by: Luca Ceresoli <luca@lucaceresoli.net>
Tested-by: Luca Ceresoli <luca@lucaceresoli.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

support/scripts: add script to test a package

This script helps in testing that a package builds fine on a wide range
of architectures and toolchains: BE/LE, 32/64-bit, musl/glibc/uclibc...

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
[yann.morin.1998@free.fr:
- completely rewrite the script from Thomas, with help from Luca
]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Luca Ceresoli <luca@lucaceresoli.net>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Acked-by: Luca Ceresoli <luca@lucaceresoli.net>
Reviewed-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Tested-by: Luca Ceresoli <luca@lucaceresoli.net>
Acked-by: Luca Ceresoli <luca@lucaceresoli.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/x11r7/xdriver_xf86-input-libinput: bump version to 0.24.0

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

configs/mx25pdk: Bump U-Boot and kernel versions

Bump U-Boot to 2017.01 version and kernel to 4.9.9.

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

package/qt5: add meaningful help texts and package URLs

Signed-off-by: Zoltan Gyarmati <mr.zoltan.gyarmati@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/qt5: remove menu comment about deprecated modules

As now the Qt version is configurable to simplify the menu display
logic we remove the menu comment about deprecated modules and
indicate it in the relevant modules' help text instead, and also
mention in the help text if a package's tech preview status has been
changed between Qt 5.6 and Qt 5.8

Signed-off-by: Zoltan Gyarmati <mr.zoltan.gyarmati@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

util-linux: link scriptreplay with libm (for isnan)

- Cherry-pick a patch from upstream to link scriptreplay to libm (for
  "isnam").
- Add a patch that improves the detection of isnan when using uClibc.
  This patch is for util-linux v2.29.1 and must be adapted to their
  master branch to be submitted upstream.

Fixes:
  http://autobuild.buildroot.net/results/2c2/2c29a78ed81ca844a87dcd076ab3e14ea080296d
  http://autobuild.buildroot.net/results/404/404b10f359b2ae8a7216729fa1bab37fed2d3d4c

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/cryptodev-linux: fix build with recent kernels

cryptodev-linux-1.8 build is broken with kernel >= 4.6.0 because of
APIs changes in the kernel. Upstream already fixed this:

https://github.com/cryptodev-linux/cryptodev-linux/commit/2b29be8ac41414ed19cb4f5d5626d9bd0d7b11a8
https://github.com/cryptodev-linux/cryptodev-linux/commit/f126e4837e6334d0464540995df7426fedf6b175
https://github.com/cryptodev-linux/cryptodev-linux/commit/f14b4706b0d04988e7e5bc8c4d2aefef9f029d9d
https://github.com/cryptodev-linux/cryptodev-linux/commit/cb186f682679383e8b5806240927903730ce85d9

Bump to latest commit to fix cryptodev-linux build.

Signed-off-by: Sébastien Szymanski <sebastien.szymanski@armadeus.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>