Ivan Sergeev [Fri, 13 Jun 2014 15:35:35 +0000 (17:35 +0200)]
systemd: add network unit file
Add and enable a systemd unit file to bring up or down network with ifup /
ifdown, analogous to the skeleton/etc/init.d/S40network init script.
Signed-off-by: Ivan Sergeev <vsergeev@kumunetworks.com>
[eric.le.bihan.dev@free.fr:
- rebase
- install service only if systemd-networkd is not selected] Signed-off-by: Eric Le Bihan <eric.le.bihan.dev@free.fr> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Gustavo Zacarias [Sat, 14 Jun 2014 10:56:48 +0000 (07:56 -0300)]
dialog: fix build failure
The old ncurses trick is no longer needed for this version, in fact it's
harmful. Switch to proper configure options.
Also disable rpath hackery since it's not required and could be
problematic. Fixes:
http://autobuild.buildroot.net/results/411/411f6171e972eab4486143dedbfd078136886ab0/
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
gcc 4.8.x does not compile a working kernel for sparc
Since the switch to 4.8.x as default, the qemu-sparc target is broken.
For a gcc bug report see here:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=60624
Switch back to gcc 4.7.x as default for sparc.
Disable 4.8/4.9 as suggested by Thomas Petazzoni.
I even disabled gcc snapshot, it works right now, because
it is an old 4.8.0 snapshot by default, but as soon as this is updated
sparc build will break.
Signed-off-by: Waldemar Brodkorb <wbx@openadk.org> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Eric Le Bihan [Fri, 13 Jun 2014 15:35:57 +0000 (17:35 +0200)]
systemd: needs kernel headers >= 3.10
Systemd needs a toolchain with kernel headers >= 3.10, as it needs
IFLA_VLAN_PROTOCOL from linux/if_link.h.
Fixes: http://autobuild.buildroot.net/results/33e/33e1447949d13bccc4076b69f902432e2131cc1c Signed-off-by: Eric Le Bihan <eric.le.bihan.dev@free.fr> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Kernel headers version 3.8.x has been deprecated since 2013.08 and thus can
be removed in 2014.08.
An automatic selection of 3.9.x headers is performed in the legacy menu.
Existing automatic selections of 3.8.x headers are modified to select
3.9.x.
As this patch removes the last occurrence of BR2_DEPRECATED_SINCE_2013_08,
the symbol is removed too.
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com> Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Peter Korsgaard [Fri, 13 Jun 2014 19:54:17 +0000 (21:54 +0200)]
haserl: bump 0.9.x version
Fixes a number of security issues and adds support for PUT/DELETE. From the
release mail:
<snip>
Stephen Röttger reported a number of security bugs, the most serious of
which is a potential heap overflow in sliding_buffer.c (file uploads).
There is a potential for remote code execution.
At the same time, I've made an *experimental* change to allow RESTful
API's possible:
* PUT and DELETE methods are handled by the POST and GET handlers.
* For mostly historical reasons, data on the URI is still called
GET.<var>, and data in the body is named POST.<var>
* If the Content-Type is not "application/x-www-form-urlencoded", Haserl
won't try to urldecode the POST contents - it will just put the body in
POST.body verbatim.
</snip>
The lua handling now uses pkg-config, so adjust the code to match.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Thomas Petazzoni [Thu, 12 Jun 2014 20:12:43 +0000 (22:12 +0200)]
util-linux: fix installation of shared libraries to not mess with pure static lib
Even when compiled with --enable-static --disable-shared, util-linux
creates some incorrect libuuid.so, libblkid.so and libmount.so
symbolic links, which confuses the compiler which thinks that a shared
library is available. This causes some build issues such as:
This commit fixes that by adding a patch to util-linux that fixes this
issue. The patch has been submitted upstream at
http://article.gmane.org/gmane.linux.utilities.util-linux-ng/9262.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Thomas Petazzoni [Fri, 13 Jun 2014 12:17:19 +0000 (14:17 +0200)]
Makefile: test if dot exists before using it in graph-depends
The 'graph-depends' logic uses the 'dot' program from Graphviz to draw
the dependency graph, but it doesn't check its existence before
starting the generation of the graph, which can lead to user confusion
as reported in:
toolchain-external: bump musl external toolchains to 1.1.1
And while we're at it, factorize the definition of the musl version,
since it's common to the definition of the tarball names for the
various supported architectures.
[Peter: Adjust Config.in info to match new version] Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Yann E. MORIN [Mon, 9 Jun 2014 22:28:56 +0000 (00:28 +0200)]
graphs/depends: do not draw transitive dependencies by default
The transitive dependencies make the graphs barely readable for large
configs, with a large number of packages.
So, just switch to not drawing the transitive dependencies by default.
By popular demand... ;-)
[Peter: reword] Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc; Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Matt Weber [Wed, 11 Jun 2014 02:19:44 +0000 (19:19 -0700)]
simicsfs: new package
[Peter: use http url as wget complains about certificate] Signed-off-by: Matt Weber <matthew.weber@rockwellcollins.com> Reviewed-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Eric Le Bihan [Wed, 11 Jun 2014 21:26:22 +0000 (23:26 +0200)]
systemd: fix AM_PATH_LIBGCRYPT not found
This patch installs a copy of libgcrypt.m4 from the libgcrypt source tarball
to systemd m4 directory.
Libgcrypt uses a custom m4 macro and not pkg-config to check if the
development files are available. Though libgcrypt support is optional in
systemd, this macro should be available whenever autoreconf is used,
otherwise the re-configuration will fail with:
configure.ac:616: warning: macro 'AM_PATH_LIBGCRYPT' not found in library
The call to autoreconf is required, as it is needed by the patch which
solves the `ln --relative` issue.
As asking the user to install the development package of libgcrypt on
the host machine or adding libgcrypt as a build dependency to systemd is
not acceptable, the required file is added to the m4 directory.
Fixes: http://autobuild.buildroot.net/results/1524d346fa17749e2ae62e063b9cfdd0de95c76a/ Fixes: http://autobuild.buildroot.net/results/10bcd92437eaa27eb61f8281c93efcb53d555e35/ Signed-off-by: Eric Le Bihan <eric.le.bihan.dev@free.fr> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
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>
liblua only uses libdl when dynamic linking is used, and certain toolchains
(E.G. bfin) doesn't provide a libdl - So only link against it if it is
needed.
At the same time change it pass the library in LIBS instead of LDFLAGS so it
ends up at the end of the linker cmdline.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Thomas Petazzoni [Wed, 11 Jun 2014 20:46:29 +0000 (22:46 +0200)]
hdparm: prevent package makefile from stripping the hdparm binary
Packages shouldn't strip executables by themselves, so that Buildroot
controls whether stripping should occur or not. This also fixes the
build on Blackfin FLAT where stripping actually doesn't work because
stripping FLAT binaries is not supported.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
infra: indent the Config.in source-ing of packages
This makes sure that a patch adding a package shows in which menu the
package is added.
Before this commit, the patch has something like this:
> diff --git a/package/Config.in b/package/Config.in
> index 7800f23..433312e 100644
> --- a/package/Config.in
> +++ b/package/Config.in
> @@ -864,6 +864,7 @@ source "package/googlefontdirectory/Config.in"
> source "package/haveged/Config.in"
> source "package/mcrypt/Config.in"
> source "package/mobile-broadband-provider-info/Config.in"
> +source "package/mypackage/Config.in"
> source "package/shared-mime-info/Config.in"
> source "package/snowball-init/Config.in"
> source "package/sound-theme-borealis/Config.in"
[> added to avoid git-am recognizing this as the patch]
After this commit, the function marker shows in which menu the new
package was added:
> diff --git a/package/Config.in b/package/Config.in
> index b1111c8..7e6e1a4 100644
> --- a/package/Config.in
> +++ b/package/Config.in
> @@ -864,6 +864,7 @@ menu "Miscellaneous"
> source "package/haveged/Config.in"
> source "package/mcrypt/Config.in"
> source "package/mobile-broadband-provider-info/Config.in"
> + source "package/mypackage/Config.in"
> source "package/shared-mime-info/Config.in"
> source "package/snowball-init/Config.in"
> source "package/sound-theme-borealis/Config.in"
To keep things consistent, this is done for Config.in.host there, even
though we don't have sub-menus there at the moment.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be> Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
supervisor: add runtime dependency on python-setuptools
supervisor has a runtime dependency on python-setuptools which was not
expressed in its Config.in file. When running supervisor without setuptools,
one gets:
Starting supervisord: Traceback (most recent call last):
File "/usr/bin/supervisord", line 5, in <module>
from pkg_resources import load_entry_point
ImportError: No module named pkg_resources
Reported-by: Sebastian Himberger <sebastian@himberger.de> Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Gustavo Zacarias [Wed, 11 Jun 2014 13:31:28 +0000 (10:31 -0300)]
wireless tools: bump to version 30.pre9
It seems there won't be any new release soon (or ever), so update to the
latest pre version for better functionality (non-ASCII ESSIDs, bigger
scan buffer for example).
Also update the bzero deprecation patch since it missed some instances.
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This variable is looked up by the Makefile throught a git command line
invocation.
Since we don't keep the '.git' folder in the packages when we clone
them, the command line fail to find anything.
Even worse, it can keep going up the tree until it finds buildroot's
'.git' directory.
Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Thomas Petazzoni [Tue, 10 Jun 2014 21:01:31 +0000 (23:01 +0200)]
znc: add patch to add missing <time.h> inclusions
With certain toolchains, znc fails to build due to missing <time.h>
inclusions. This commit fixes that by adding a patch that adds those
missing inclusions.
Gustavo Zacarias [Wed, 11 Jun 2014 10:39:39 +0000 (07:39 -0300)]
dhcp: bump to version 4.1-ESV-R9
Also install the SysV initscripts from the proper hook.
Since the scripts check for the proper files to be present there's no
need to exclude them according to options since they won't try to start
anything.
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
perl-cross: bump to version 0.8.5 and move to github
The perl-cross tarball allows to patch the Perl distribution (in a
_POST_EXTRACT_HOOKS). The github helper gives a copy of the project which
allows to build this tarball. So, the github helper can not be used here.
[Peter: mention the change to github] Signed-off-by: Francois Perrad <francois.perrad@gadz.org> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
pkg-perl: fix for Perl XS packages configured by Makefile.PL
ExtUtils::MakeMaker adds all the header files used by the perl as
dependencies to the generated Makefile. This means that the generated
Makefile will depend on the system's header files.
Usually this is not a problem, because when building the target package,
these header files will indeed be found in $(STAGING_DIR). However, some
distro's add an extra header file to the system's perl. This header is
also included in the generated Makefile, which makes the build fail
because it doesn't exist in $(STAGING_DIR).
As a work-around, explicitly create this header file in $(STAGING_DIR).
It doesn't hurt to create it even if the system perl doesn't need it.
Gustavo Zacarias [Tue, 10 Jun 2014 10:18:25 +0000 (07:18 -0300)]
bind: security bump to version 9.9.5
Fixes CVE-2014-0591 (a crafted query against an NSEC3-signed zone can
crash BIND).
The 9.9.x series is the new ESV vesion, 9.6.x has been retired.
Also cleanup the initscript while at it.
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This commit significantly reduces the build time of host-gettext and
gettext, by using the capacity of gettext to handle build things in a
certain subdirectory:
- For the host variant of gettext, we only need the gettext-tools,
available in the directory of the same name in the gettext sources.
- For the target variant of gettext, we only need the gettext library
libintl, available in the gettext-runtime directory in the gettext
sources.
So by using appropriate values of GETTEXT_SUBDIR and
HOST_GETTEXT_SUBDIR, we only build what's necessary. Moreover, by
manually patching gettext-tools/Makefile.in and
gettext-runtime/Makefile.in, we make sure to not build and install
things like examples, documentation and so on.
In addition to this, these changes avoid the need to autoreconfigure
the gettext package, which was particularly long.
Thanks to these changes, the build time of gettext goes from 1 minutes
and 37 seconds to just 24 seconds, and the build of host-gettext goes
from 2 minutes and 18 seconds to 1 minute and 13 seconds.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Tested-by: Andrew Ruder <andrew.ruder@elecsyscorp.com> Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr> Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
gettext: remove support for gettext-tools on target
This commit removes the BR2_PACKAGE_GETTEXT_TOOLS option, which could
be used to install gettext tools on the target. This is not needed,
because Buildroot is not designed to provide a full development
environment on the target, and gettext translation files should be
processed on the build machine, using the host gettext tools.
Remove this option will allow to optimize the build time of gettext on
the target, by only building the gettext runtime libraries.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Tested-by: Andrew Ruder <andrew.ruder@elecsyscorp.com> Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Commit 9ff640ae30ca588228030faf2c6fd13af292cb66 ("minidlna: needs
gettext and host-gettext tools if locale is enabled") modified the
minidlna package to fix gettext related issues. As part of this patch,
a select of BR2_PACKAGE_GETTEXT_TOOLS was added, which according to
the commit title is used to have host-gettext tools installed.
However, this is not what this option is about: this option is about
having gettext tools installed on the target.
Since this is not what minidlna needs, and we anyway plan to remove
this BR2_PACKAGE_GETTEXT_TOOLS option, this commit removes this
incorrect select statement.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Tested-by: Andrew Ruder <andrew.ruder@elecsyscorp.com> Cc: Bernd Kuhls <berndkuhls@hotmail.com> Tested-by: Bernd Kuhls <bernd.kuhls@t-online.de> Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
toolchain-external: support only one Linaro AArch64 toolchain at a time
Linaro toolchains are released so frequently (every month) that it
doesn't make much sense to support 3 consecutive versions. So, like we
do for ARM big-endian, let's support only one version at a time.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
toolchain-external: support only one Linaro ARM toolchain at a time
Linaro toolchains are released so frequently (every month) that it
doesn't make much sense to support 3 consecutive versions. So, like we
do for ARM big-endian, let's support only one version at a time.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
toolchain-external: do not pass TOOLCHAIN_EXTERNAL_CFLAGS for kernel headers check
When getting the sysroot used for the kernel headers version check,
passing TOOLCHAIN_EXTERNAL_CFLAGS causes a problem when used with
multilib toolchains, where only the main sysroot has the header files,
and the other sysroots only have the libraries.
Since the kernel headers version used is normally the same for all
sysroots, this commit solves this problem by removing the
TOOLCHAIN_EXTERNAL_CFLAGS argument when calling
toolchain_find_sysroot, so that it returns the main sysroot, in which
<linux/version.h> can be found for the kernel headers version check.
Cc: Yann E. MORIN <yann.morin.1998@free.fr> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
toolchain-external: apply kernel headers check to non-custom toolchains
Since the introduction of the kernel headers Config.in options, the
external toolchain logic had a check for custom external toolchains to
verify that the kernel headers version entered by the user matches the
one of the toolchain. However, this check was not made for non-custom
external toolchains (i.e the built-in profiles, such as Linaro,
CodeSourcery and al.), making the assumption that the Buildroot
developers will do the right selection.
However, it is quite nice when bumping external toolchains to have
this automatic kernel headers version check, to ensure we select the
appropriate kernel headers version.
Therefore, this commit makes the kernel headers version check
applicable to non-custom external toolchains.
Cc: Yann E. MORIN <yann.morin.1998@free.fr> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
[Thomas:
- fix license version: the license is GPLv2+, not GPLv2
- make sure the ncurses backend is selected if none of the other
backends are selected.
- add dependency on libiconv of the curses and gtk2 backends, needed
in !locale configurations.]
Phil Eichinger [Thu, 29 May 2014 15:05:33 +0000 (17:05 +0200)]
sispmctl: new package
[Thomas:
- Add missing select of BR2_PACKAGE_LIBUSB, since
BR2_PACKAGE_LIBUSB_COMPAT depends on it
- Add missing thread dependency, since libusb requires threads.
- Add missing wide-char dependency. Without wide char support,
sispmctl doesn't build.
- Rewrap the Config.in help text, and remove trailing whitespaces.
- License is GPLv2+, not GPLv2.
- Remove --disable-dependency-tracking from <pkg>_CONF_OPT. That's a
global, standard, autoconf option, and there's no reason to pass it
at the per-package level.]
Signed-off-by: Phil Eichinger <phil@zankapfel.net> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
linux-zigbee is a mix of utility tools that implements a 802.15.4 stack
along with a usefull library for any program that would need to use this
stack.
Note that the name is mislead, for licensing reason, this does *NOT*
implement the ZigBee protocol.
[Thomas:
- add dependency on threads, propagated from libnl
- improve the Config.in description by borrowing more text from the
upstream website
- fix the prompt of sub-options to be more consistent, and add help
texts where appropriate
- fix indentation of the BR2_PACKAGE_LINUX_ZIGBEE_TESTS option help
text
- add missing dependencies on host-pkgconf, host-flex and host-bison
- add missing SoB line in the patch disabling test-serial. Maxime is
a well-known contributor, so I assumed we had his SoB.
- fix indentation of <pkg>_CONF_OPT.
- remove <pkg>_INSTALL_STAGING = YES since the package does not
install any library, and the two headers it installs are available
through the toolchain kernel headers.
- add comment to explain <pkg>_AUTORECONF = YES.]
Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Bump zedboard defconfig to version 2014.1. This makes it possible to use
uboot-spl instead of the xilinx specific fsbl. However, the result is kinda
hacky due to the licensing state of ps7_init.{c,h} needed for building uboot
spl. Directions for building a proper boot.bin using the new methodolgy is
included in the readme.txt.
Signed-off-by: Spenser Gilliland <spenser@gillilanding.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Gustavo Zacarias [Tue, 20 May 2014 13:44:22 +0000 (10:44 -0300)]
ruby: remove rubygems
It's currently in a non-working state since it requires a couple of
extensions that don't build at the moment (they try to execute tests in
configure) and also requires a target compiler.
So remove it to avoid false expectations and reclaim target space back
of about 1.5 MiB.
[Thomas: slightly reword comments.]
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Steve Thomas [Mon, 19 May 2014 19:48:13 +0000 (19:48 +0000)]
gcc: add a BR2_GCC_ENABLE_GRAPHITE option
The gcc graphite optimisations such as loop-interchange, blocking
and loop-flattening, also known as graphite are an optional feature of
gcc that is very well supported since about gcc version 4.5.
This patch adds support for graphite for the toolchain as an optional
flag for versions 4.8 onwards as an optional flag, that is disabled by
default.
Signed-off-by: Steve Thomas <scjthm@live.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Steve Thomas [Mon, 19 May 2014 19:48:12 +0000 (19:48 +0000)]
cloog: new package
CLooG is a free software and library to generate code for scanning
Z-polyhedra.
cloog is needed for the optional graphite optimisations that are
supported by gcc since version 4.5. Therefore this package is required
for the toolchain to support graphite.
Graphite optimisations primarily involve loop blocking flattening and
interchage so are probably of mimimal use in an embedded system where
small sizes are favoured.
cloog depends on isl.
[Thomas:
- Add patch to add missing CMake related files to the release
tarball, preventing the build from succeeding.
- bump to 0.18.2
- disable libtool patch, which doesn't apply]
Signed-off-by: Steve Thomas <scjthm@live.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Steve Thomas [Mon, 19 May 2014 19:48:11 +0000 (19:48 +0000)]
isl: new package
isl is a library for manipulating sets and relations of integer points
bounded by linear constraints.
isl is needed for the optional graphite optimisations that are supported
by gcc since version 4.5. Therefore this package is required for the
toolchain to support graphite.
Graphite optimisations primarily involve loop blocking flattening and
interchage so are probably of mimimal use in an embedded system where
small sizes are favoured.
[Thomas:
- bump to 0.12.2 (cannot use 0.13, incompatible with cloog 0.18.2,
comment added about this)
- use .bz2 tarball
- disable libtool patch, which doesn't apply]
Signed-off-by: Steve Thomas <scjthm@live.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Yann E. MORIN [Sun, 8 Jun 2014 20:02:23 +0000 (22:02 +0200)]
filesystems: also chown symlinks
Currently, the symlinks in the generated filesystems will have the
UID of the user running the build, because 'chown' does not change
the ownership of symlinks, by default.
Although the implications are limited, some may not want that UID
to leak in the generated filesystems.
So, use 'chown -h' so even symlinks get properly chowned.
Reported-by: Angelo Dureghello <angelo@barix.com> Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr> Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
James Hogan [Mon, 9 Dec 2013 17:05:04 +0000 (17:05 +0000)]
uclibc: add patches to fix MIPS __SIGRTMAX and SIG127
Add two uClibc 0.9.33.2 patches which fix issues relating to high signal
number handling on MIPS.
The first patch (0056) fixes _NSIG (and as a result __SIGRTMAX) to match
glibc. This fixes GDB on MIPS with uClibc, which cannot handle
__SIGRTMAX == 128 and emits the error:
GDB bug: target.c (gdb_signal_from_host): unrecognized real-time signal
This patch is from uClibc commit 2da958760f79 (MIPS: set _NSIG to
128, not 129. This matches glibc.) and applies without conflicts.
The second patch (0057) fixes the wait status macros to correctly
interpret status 0x007f on MIPS (other arches don't have signal 127).
This patch is from uClibc commit 4a96b9486871 (bits/waitstatus.h:
correctly interpret status 0x007f on MIPS) and applies without
conflicts.
Signed-off-by: James Hogan <james.hogan@imgtec.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
James Hogan [Mon, 9 Dec 2013 17:05:03 +0000 (17:05 +0000)]
uclibc: add patch to fix arch specific eventfd defs
Add uClibc 0.9.33.2 patch to fix arch specific eventfd definitions,
particularly EFD_NONBLOCK. The definition in sys/eventfd.h was 04000
(0x800), however MIPS uses 0200 (0x80). This resulted in QEMU built for
MIPS hanging at various points until input is received due to a blocking
poll of stdin.
The patch is a backport of uClibc commit fd355bc1dbcb (eventfd.h: Use
new "bits/" scheme for arch-specific flags). The only conflict was minor
in Makefile.in.
Signed-off-by: James Hogan <james.hogan@imgtec.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Peter Kümmel [Sun, 1 Jun 2014 10:18:28 +0000 (12:18 +0200)]
sunxi-mali: use only -I${includedir} as pkg-config Cflags
Using only -I${includedir} in common practice.
Settings only -I${includedir}/GLES2 breaks Qt5's CMake files,
because at configure time CMake only searchs the paths reported
by pkg-config, and not even /usr/include is used as default.
Even though pkg-config strips out standard include path, that's not
the case with pkgconf (which we are using) in cross-compilation,
which correctly reports the /usr/include dir prefixed with the
sysroot.
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr> Signed-off-by: Peter Kümmel <syntheticpp@gmx.net> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>