In 36568732e4, we expanded toolchain.cmake to also define the value for
CMAKE_SYSTEM_VERSION, as the cmake documentation states that it must be
manually defined when doing cross-compilation [0]:
When the CMAKE_SYSTEM_NAME variable is set explicitly to enable
cross compiling then the value of CMAKE_SYSTEM_VERSION must also
be set explicitly to specify the target system version.
However, the fix in 36568732e4 uses the version of the kernel headers,
assuming that would be the oldest kernel we could run on. Yet, this is
not the case, because glibc (for example) has fallbacks to support
running on kernels older than the headers it was built against.
The cmake official wiki [1] additionally states:
* CMAKE_SYSTEM_VERSION : optional, version of your target system, not
used very much.
Folllowed a little bit below, by:
* CMAKE_TOOLCHAIN_FILE : absolute or relative path to a cmake script
which sets up all the toolchain related variables mentioned above
For instance for crosscompiling from Linux to Embedded Linux on PowerPC
this file could look like this:
# this one is important
SET(CMAKE_SYSTEM_NAME Linux)
#this one not so much
SET(CMAKE_SYSTEM_VERSION 1)
[...]
Furthermore, using the kernel headers version can be a bit misleading (as
it really looks like is is the correct version to use when it is not),
while it is obvious that 1 is not really the output of `uname -r` and
thus is definitely not misleading.
Finally, random searches [2] about CMAKE_SYSTEM_VERSION, mostly only
turns up issues related with Windows, Mac-OS, and to a lesser extent,
Android (where it is forcibly set to 1), with issues realted to running
under just Linux (as opposed to Adnroid) mostly non-existent.
Consequently, we revert to using the value that is suggested in the
cmake WiKi, i.e. 1, and which is basically what we also used as a
workaround in the azure-iot-sdk-c paclkage up until d300b1d3b1.
A case were we will need to have a real kernel version, is if we one day
have a cmake-based pacakge that builds and installs a kernel module [3],
because it will need the _running_ kernel version to install it in
/lib/modules/VERSION/, but in that case it will anyway most probably
not be the headers version.
[0] https://cmake.org/cmake/help/v3.8/variable/CMAKE_SYSTEM_VERSION.html
[1] https://gitlab.kitware.com/cmake/community/wikis/doc/cmake/CrossCompiling
[2] https://duckduckgo.com/?q=CMAKE_SYSTEM_VERSION
[3] https://stackoverflow.com/questions/38205745/cmake-system-version-not-updated-for-new-kernel
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Samuel Martin <s.martin49@gmail.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit fc8a5f56b9)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Quoting the CMake documentation:
When the CMAKE_SYSTEM_NAME variable is set explicitly to enable cross
compiling then the value of CMAKE_SYSTEM_VERSION must also be set
explicitly to specify the target system version.
Thus, we should also set CMAKE_SYSTEM_VERSION in toolchainfile.cmake. It
is supposed to be set to the value of `uname -r` on the target. We don't
have that exact value available (unless we build the kernel), but the
value of BR2_TOOLCHAIN_HEADERS_AT_LEAST contains the (minimum) version
of the kernel it will run on, so it should be OK for all practical
purposes.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Cc: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 36568732e4)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Since the bump to version 1.15, libsquish can conditionally
build/install its shared library, which makes it possible to re-enable
this package for BR2_STATIC_LIBS configurations.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit ae4b9e4afb)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
While testing the new supertuxkart 0.10-beta1 version, the build
failed due to missing libsquash shared library.
Indeed, by default (see config file) the shared library is not build
and all symlinks created by LIBSQUISH_INSTALL_{STAGING,TARGET}_CMDS
are curently broken.
To fix that, add USE_SHARED=1 to LIBSQUISH_MAKE_ENV.
Also add a patch to allow reinstall the package.
Signed-off-by: Romain Naour <romain.naour@gmail.com>
Cc: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit b621273755)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
releases.nixos.org gives some weird XML pages. The upstream location is
nixos.org/releases.
This side uses HSTS, so switch to https to avoid a needless redirect.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit e2fa462426)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Commit ec69d3820f (package/lighttpd: bump to version 1.4.52) bumped the
lighttpd version but forgot to adjust the upstream hash comment.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit b99b09caee)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 78f0c1fe4a)
[Peter: drop 4.19.x/4.20.x, linux / hash changes]
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
As xenomai release tarballs ship with a configure script, there is only a
need for autoreconfiguring when patches are applied.
The last patch was removed with git commit
de993bc23a but the AUTORECONF line remained.
Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 4219c58c38)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Upstream CHANGELOG entry for 4.3.0 lists these fixes:
- CVE-2018-18408 use-after-free in post_args (#489)
- CVE-2018-18407 heap-buffer-overflow csum_replace4 (#488)
- CVE-2018-17974 heap-buffer-overflow dlt_en10mb_encode (#486)
- CVE-2018-17580 heap-buffer-overflow fast_edit_packet (#485)
- CVE-2018-17582 heap-buffer-overflow in get_next_packet (#484)
- CVE-2018-13112 heap-buffer-overflow in get_l2len (#477 dup #408)
Drop tr_cv_libpcap_version and ac_cv_have_bpf; unused in current
configure script.
Make configure script use pcap-config to list library dependencies.
Unfortunately, pcap-config is not entirely correct, so we still need to
set the LIBS variable for static linking.
Use the smaller tar.xz archive.
Add license file hash.
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit f5961ff56a)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Commit 62d5558f76 (utils/get-developers: make it callable from elsewhere
than the toplevel directory) tried to fix this by passing in the toplevel
directory when the DEVELOPERS file is parsed.
Unfortunately this is not enough, as E.G. also the paths listed in the
patches are relative to the toplevel directory, causing it to not match the
entries in the DEVELOPERS file.
In concept this can be fixed by also passing the toplevel directory to the
Developers class, but the simplest solution is just to chdir to the toplevel
Buildroot directory before calling any of the getdeveloperlib functions.
This does require us to finish parsing command line arguments (which opens
the provided patch files) to not get into trouble with relative paths to
patches before chdir'ing / initializing getdeveloperlib.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 45aabcddc5)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This reverts commit 62d5558f76.
This actually does not work, as patches contain paths relative to the
toplevel directory as well.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 74fbc20504)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes regressions introduced by the v8.14.0 security release. From the
announcement:
The 8.14.0 security release introduced some unexpected breakages on the 8.x
release line. This is a special release to fix a regression in the HTTP
binary upgrade response body and add a missing CLI flag to adjust the max
header size of the http parser.
https://github.com/nodejs/node/blob/master/doc/changelogs/CHANGELOG_V8.md#8.15.0
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 6f68e5cf72)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Explain how this can be used with git send-email, and how git can be
configured to use it automatically.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 507d56a6be)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The Config.in comment of lua-msgpack-native only indicates that it
needs Lua 5.1, while the package can actually build with LuaJIT. This
commit adjusts the comment to match the reality.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 6141506760)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Before this commit, the grub configuration file was copied to the
TARGET_DIR in a post-image hook, after the filesystem has been
generated. It was kinda working because the board/pc's grub
configuration and the default one are the same and the later was
copied during the build process of the grub2 package.
This commit ensures the custom board/pc grub configuration is copied at
the right time.
Signed-off-by: Grégoire Delattre <gregoire.delattre@gmail.com>
Reviewed-by: Matt Weber <matthew.weber@rockwellcollin.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 0b28397a9b)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The following vulnerabilities have been fixed:
- wnpa-sec-2018-51 The Wireshark dissection engine could crash. Bug 14466. CVE-2018-19625.
- wnpa-sec-2018-52 The DCOM dissector could crash. Bug 15130. CVE-2018-19626.
- wnpa-sec-2018-53 The LBMPDM dissector could crash. Bug 15132. CVE-2018-19623.
- wnpa-sec-2018-54 The MMSE dissector could go into an infinite loop. Bug 15250. CVE-2018-19622.
- wnpa-sec-2018-55 The IxVeriWave file parser could crash. Bug 15279. CVE-2018-19627.
- wnpa-sec-2018-56 The PVFS dissector could crash. Bug 15280. CVE-2018-19624.
- wnpa-sec-2018-57 The ZigBee ZCL dissector could crash. Bug 15281. CVE-2018-19628.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 2ec38321d6)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
- Remove first patch (inet_pton test removed since version 2.5.0 and
12d63c428f)
- Remove second patch (already in version, since
3761ef3785)
- Remove WIRESHARK_AUTORECONF=yes as configure.ac is not patched anymore
- gcrypt is a mandatory dependency since version 2.4.0 and
6b84ddee83
- geoip support has been replaced by maxminddb since version 2.5.0 and
a1da75c554
- Drop includedir "hack", this is not needed anymore and wireshark is
now installing files in target. With this hack, build fails on:
wireshark: installs files in /home/fabrice/buildroot/output/target//home/fabrice/buildroot/output
package/pkg-generic.mk:310: recipe for target '/home/fabrice/buildroot/output/build/wireshark-2.6.4/.stamp_target_installed' failed
- Disable all new optional features
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
[Arnout: keep optional dependencies alphabetic on Config.in name;
add comment of libgcrypt -> libgpg-error dependency.]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
(cherry picked from commit 72bafd5e45)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
- lua 5.3 or luajit is not supported
- Don't specify the path to find lua as pkgconfig is used for lua
(this is not the case for the other options)
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit fb59a8857e)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
In order to provide info about a bundle file, e.g. running "rauc info
bundle.raucb", rauc needs to use the unsquashfs progam from the squashfs
package.
This was not documented until upstream commit 10c501c12752 ("docs:
integration: document need for 'unsquashfs' tool for 'rauc info'"), but
is already present in rauc.
Signed-off-by: Trent Piepho <tpiepho@impinj.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 4a4dba21b9)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
sqlite3 refuses to be built with -ffast-math (a side effect of -Ofast) when it
falls back to implementing its own isnan() function.
sqlite3.c: In function ‘sqlite3IsNaN’:
sqlite3.c:28554:3: error: #error SQLite will not work correctly with the -ffast-math option of GCC.
To work around this, when -Ofast is used replace with -O3.
Signed-off-by: Joshua Henderson <joshua.henderson@microchip.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 83781f11dc)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Just like the target acpica package needs host-flex and host-bison,
the host variant also needs the same dependencies. This allows to fix
the build of "make host-acpica", which was detected thanks to
per-package directory support.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 8681430628)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Since commit aa556e2035
("utils/genrandconfig: test with BR2_OPTIMIZE_2=y"), we are doing
builds at -O2 instead of -Os. This has unveiled an issue in the
strsep.c file:
strsep.c:65:23: error: register name not specified for 'delim'
register const char *delim;
This strsep.c compatibility code is compiled in if HAVE_STRSEP is not
defined, but dhcpdump does not use any kind of configure script to
detect the availability of strsep(). Therefore by default, it gets
compiled in, and the "register" specifier used for some variable
declarations in strsep.c cause build issues at -O2.
A previous commit in Buildroot from
c2a7f0d605 ("dhcpdump: Fix strsep()
feature test"), attempted to fix this problem by changing the test on
HAVE_STRSEP by a test on _BSD_SOURCE.
Unfortunately, _BSD_SOURCE is not meant to be tested: it's a feature
macro that is meant to be *defined* by some code to tell the C library
headers to expose (or not) some given functionality.
So instead, we basically revert commit
c2a7f0d605 by dropping the patch, and
pass -DHAVE_STRSEP in the CFLAGS when building dhcpdump.
Fixes:
http://autobuild.buildroot.net/results/7231170d3d3e3637f02382c1a0a96009b0527618/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 136c8862cf)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Building a minimal defconfig such as:
BR2_arm=y
BR2_TOOLCHAIN_EXTERNAL=y
BR2_TOOLCHAIN_EXTERNAL_CUSTOM=y
BR2_TOOLCHAIN_EXTERNAL_DOWNLOAD=y
BR2_TOOLCHAIN_EXTERNAL_URL="http://autobuild.buildroot.org/toolchains/tarballs/br-arm-full-2018.05.tar.bz2"
BR2_TOOLCHAIN_EXTERNAL_GCC_4_9=y
BR2_TOOLCHAIN_EXTERNAL_HEADERS_4_1=y
BR2_TOOLCHAIN_EXTERNAL_LOCALE=y
BR2_TOOLCHAIN_EXTERNAL_CXX=y
BR2_INIT_NONE=y
BR2_SYSTEM_BIN_SH_NONE=y
BR2_PACKAGE_XORG7=y
BR2_PACKAGE_XAPP_RGB=y
by running "make xapp_rgb" gives the following build failure:
checking for RGB... configure: error: in `/home/test/buildroot/output/build/xapp_rgb-1.0.6':
configure: error: The pkg-config script could not be found or is too old. Make sure it
is in your PATH or set the PKG_CONFIG environment variable to the full
path to pkg-config.
Alternatively, you may set the environment variables RGB_CFLAGS
and RGB_LIBS to avoid the need to call pkg-config.
See the pkg-config man page for more details.
The configure script uses pkg-config, but host-pkgconf is missing in
the list of dependencies.
This issue was detected thanks to per-package directory support.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 401b7b94a3)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The first stage bootloader on LEGO MINDSTORMS EV3 only loads the first
256K of the flash memory (this is hard-coded in an EEPROM). So, the
second stage bootloader (U-Boot) cannot exceed this size.
Signed-off-by: David Lechner <david@lechnology.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit b2e5014af5)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
