In case the git backend gets killed right in-between it finished
initialising the repository, but before it could add the remote,
we'd end up with a repository without the 'origin' remote, so we
would not be able to change its URL.
Another case that may happen (like in the build failure, below),
is that the repository was initialised with a previous version
of Buildroot, before the commit e17719264b (download/git: don't
require too-recent git) was applied, and that trepository was
still lying around...
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr> Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com> Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com> Cc: Peter Korsgaard <peter@korsgaard.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Erik Larsson [Sat, 10 Mar 2018 11:51:40 +0000 (12:51 +0100)]
freescale-imx: add config option for i.MX8M
Signed-off-by: Erik Larsson <karl.erik.larsson@gmail.com> Signed-off-by: Christopher Dahlberg <crille.dahlberg@gmail.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Erik Larsson [Sat, 10 Mar 2018 11:51:39 +0000 (12:51 +0100)]
imx-mkimage: new package
This package add support for imx-mkimage for iMX8M.
Currently the only hardware it's tested for.
Signed-off-by: Erik Larsson <karl.erik.larsson@gmail.com> Signed-off-by: Christopher Dahlberg <crille.dahlberg@gmail.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
output/build/llvm-5.0.1/include/llvm/Support/ConvertUTF.h:203:53:
error: ‘std::wstring’ has not been declared
bool ConvertUTF8toWide(llvm::StringRef Source, std::wstring &Result);
Also propagate dependency to mesa3d
Signed-off-by: Valentin Korenblit <valentin.korenblit@smile.fr> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
James Byrne [Mon, 9 Apr 2018 16:08:01 +0000 (17:08 +0100)]
Makefile: avoid executing 'git log' each time SOURCE_DATE_EPOCH is used
If SOURCE_DATE_EPOCH is not defined it was given a definition that
caused 'git log' to be executed each time the variable is referenced,
which is not very efficient given that the answer cannot change.
This commit moves the definition of BR2_VERSION_GIT_EPOCH after the
inclusion of Makefile.in (so that GIT is defined) and makes it a
simply expanded variable so that it is only evaluated once.
Signed-off-by: James Byrne <james.byrne@origamienergy.com> Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Fixes CVE-2018-1000156: arbitrary command execution in ed-style patches.
Depend on MMU for now, because the patch adds a fork() call. Upstream
later switched to gnulib provided execute(), so this dependency can be
dropped on the next version bump.
Signed-off-by: Baruch Siach <baruch@tkos.co.il> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Chris Brandt [Sat, 7 Apr 2018 13:35:43 +0000 (08:35 -0500)]
fs/cramfs: add xip option
Cramfs now has the ability to XIP all ELF files.
Cc: Nicolas Pitre <nico@linaro.org> Signed-off-by: Chris Brandt <chris.brandt@renesas.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Chris Brandt [Sat, 7 Apr 2018 13:35:42 +0000 (08:35 -0500)]
package/cramfs: change to new site location
As of Linux-4.15, cramfs now has a official maintainer again.
Additionally, that person is hosting and maintaining a new version of
cramfs-tools.
Patches 0001-endian.patch and 0003-fix-missing-types.patch are no longer
needed because they have been upstreamed.
However, since patch 0002-cygwin_IO.patch is so old, it is being
removed until someone that needs it can reapply and test it. At
that point, they should submit the changes to the new cramfs-tools
maintainer instead of adding the patches back here.
Please note that cross-endian support for cramfsck is not supported
at this time.
Cc: Nicolas Pitre <nico@linaro.org> Signed-off-by: Chris Brandt <chris.brandt@renesas.com> Reviewed-by: Matt Weber <matthew.weber@rockwellcollins.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Passing the Buildroot make variable UCLIBC_EXTRA_CFLAGS into the
UCLIBC_EXTRA_CFLAGS fed into uClibc's make environment therefore no
longer makes sense. Therefore, just pass $(TARGET_ABI).
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
The HOST_GCC_COMMON_WRAPPER_TARGET_FLOAT_ABI variable was incorrectly
used as HOST_GCC_COMMON_WRAPPER_TARGET_FLOATABI_, which would always
be empty.
In practice, this block of code is only used on Blackfin (which has
BR2_GCC_ARCH_HAS_CONFIGURABLE_DEFAULTS unset) and Blackfin doesn't use
float-abi, so this block of code is in fact useless. But even if it's
useless, it's better to have it without typos.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
This variable was added as part of an EGL i.MX6 workaround in commit
(7e9f459adf), but appears to currently be abandonded and unused. Remove the use
of the variable.
Signed-off-by: Joshua Henderson <joshua.henderson@microchip.com> Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
libbsd: add upstream patch to fix musl build issue
Commit d879be3049598f35d7b959784e1a6c2811ba2036 enabled the build for non-glibc
toolchains and added a patch from upstream to fix one remaining build issue
with musl.
The upstream patch defines the glibc specific macro `__GLIBC_PREREQ` in
`include/bsd/sys/cdefs.h`, but the patch does not include this header
file in `include/bsd/stdlib.h` which uses the macro.
Upstream did committed an earlier patch which adds this header inclusion
to `include/bsd/stdlib.h`. So, add this upstream patch as well to fix a
build error when building the hcitop tool from the not yet added package
bluez-alsa.
Note, that the upstream patch address the issue that `<sys/cdefs.h>` is
a non-portable header. However, this is not the issue which needs to be
fixed by this patch, but solely the needed inclusion of the
`include/bsd/stdlib.h` header file, to see the definition of
`__GLIBC_PREREQ`.
Where _estrdup is the actual function implemented by the PHP core. If
this header file is not included, and some code uses estrdup, one ends
up with an undefined reference. This happens when libexpat support is
enabled. This commit adds a PHP patch that fixes this issue. The patch
has been submitted upstream through a Github pull request.
Commit 2759ba03979 (liblo: disable werror to fix build issues) tried to
fix the issue by passing --disable-werror, but this is not a recognised
option.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr> Cc: Alex Baldwin <alexbaldwinmusic@gmail.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Alistair Francis <alistair@alistair23.me> Tested-by: Matt Weber <matthew.weber@rockwellcollins.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
André Hentschel [Sat, 7 Apr 2018 12:59:03 +0000 (14:59 +0200)]
wireshark: bump version to 2.2.14 (security)
Security fixes since 2.2.12:
- wnpa-sec-2018-15
The MP4 dissector could crash. (Bug 13777)
- wnpa-sec-2018-16
The ADB dissector could crash. (Bug 14460)
- wnpa-sec-2018-17
The IEEE 802.15.4 dissector could crash. (Bug 14468)
- wnpa-sec-2018-18
The NBAP dissector could crash. (Bug 14471)
- wnpa-sec-2018-19
The VLAN dissector could crash. (Bug 14469)
- wnpa-sec-2018-20
The LWAPP dissector could crash. (Bug 14467)
- wnpa-sec-2018-23
The Kerberos dissector could crash. (Bug 14576)
- wnpa-sec-2018-05
The IEEE 802.11 dissector could crash. Bug 14442, CVE-2018-7335
- wnpa-sec-2018-06
Multiple dissectors could go into large infinite loops. All ASN.1 BER dissectors (Bug 14444), along with the DICOM (Bug 14411), DMP (Bug 14408), LLTD (Bug 14419), OpenFlow (Bug 14420), RELOAD (Bug 14445), RPCoRDMA (Bug 14449), RPKI-Router (Bug 14414), S7COMM (Bug 14423), SCCP (Bug 14413), Thread (Bug 14428), Thrift (Bug 14379), USB (Bug 14421), and WCCP (Bug 14412) dissectors were susceptible.
- wnpa-sec-2018-07
The UMTS MAC dissector could crash. Bug 14339, CVE-2018-7334
- wnpa-sec-2018-09
The FCP dissector could crash. Bug 14374, CVE-2018-7336
- wnpa-sec-2018-10
The SIGCOMP dissector could crash. Bug 14398, CVE-2018-7320
- wnpa-sec-2018-11
The pcapng file parser could crash. Bug 14403, CVE-2018-7420
- wnpa-sec-2018-12
The IPMI dissector could crash. Bug 14409, CVE-2018-7417
- wnpa-sec-2018-13
The SIGCOMP dissector could crash. Bug 14410, CVE-2018-7418
- wnpa-sec-2018-14
The NBAP disssector could crash. Bug 14443, CVE-2018-7419
Reported-by: André Hentschel <nerv@dawncrow.de> Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr> Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com> Cc: André Hentschel <nerv@dawncrow.de> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
When a package contains a relative symlink which first component is '..'
(thus pointing one directory higher), for example package 'meh' contains
this symlink:
foo/bar -> ../buz
then it would be stored as 'meh-version./buz' because of the
transform-name pattern replacement.
Fix it to only match the leading './'.
Reported-by: Ricardo Martincoski <ricardo.martincoski@gmail.com> Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr> Cc: Ricardo Martincoski <ricardo.martincoski@gmail.com> Cc: Peter Korsgaard <peter@korsgaard.com> Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com> Reviewed-by: Ricardo Martincoski <ricardo.martincoski@gmail.com> Tested-by: Ricardo Martincoski <ricardo.martincoski@gmail.com> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Update the ZEROMQ_SITE with the active repository.
Remove the two patches because they were both applied on zeromq.
As a result, the ZEROMQ_AUTORECONF was removed.
Signed-off-by: Asaf Kahlon <asafka7@gmail.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Since the bump of libglib2 to 2.56.0, it fails to build on uClibc with:
gthreadedresolver.c: In function 'do_lookup_records':
gthreadedresolver.c:875:16: error: 'res' undeclared (first use in this function); did you mean '_res'?
res_nclose (&res);
^~~
_res
This commit adds an upstream patch that fixes this problem.
This is a required component to complete the configuration. It also makes sense
to add ubifs so that this board can be flashed with the board/atmel/flasher.sh
script.
Signed-off-by: Joshua Henderson <joshua.henderson@microchip.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Romain Naour <romain.naour@gmail.com>
[Thomas: mention the libunwind dependency as being needed for 'strace
-k' in the Config.in help text, as suggested by Thomas De
Schampheleire.] Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Martin Bark [Sat, 24 Mar 2018 14:10:15 +0000 (14:10 +0000)]
package/rpi-bt-firmware: change to LibreELEC github repo
Include BCM4345C0.hcd for the rpi3 b+
Signed-off-by: Martin Bark <martin@barkynet.com>
[Thomas: rebase on master, use make foreach loop instead of shell for
loop.] Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Martin Bark [Sat, 24 Mar 2018 14:10:14 +0000 (14:10 +0000)]
package/rpi-firmware: improve installation of dtb files
Don't list specific dtb files, instead install all dtb files in the same
way as dtbo files are currently done.
Signed-off-by: Martin Bark <martin@barkynet.com>
[Thomas: use make foreach loop instead of shell for loop.] Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Martin Bark [Sat, 24 Mar 2018 14:10:12 +0000 (14:10 +0000)]
configs/raspberrypi*: bump kernel version
Now based on 4.14.29 (from 4.9.79)
Signed-off-by: Martin Bark <martin@barkynet.com>
[Thomas: adjusted on top of Yann E. Morin changes to use a tarball
instead of a git clone.] Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
This package contains a kernel module from Intel, which could only be
used together with Intel userspace tools provided under NDA, which
also come with the same kernel module. The copy of the kernel module
available on SourceForge is only provided to comply with the GPLv2
requirement. Intel engineers were even surprised it even built and
were not willing to make any effort to fix their tarball naming to
contain a version number. Therefore, it does not make sense for
Buildroot to provide such a package.
See https://sourceforge.net/p/e1000/bugs/589/ for the discussion.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com> Acked-by: Carlos Santos <casantos@datacom.ind.br> Acked-by: Romain Naour <romain.naour@gmail.com>
[Thomas:
- remove double "only" noticed by Carlos
- fix typo packaged -> package noticed by Romain
- fix check-package warning noticed by Ricardo] Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
board/atmel: use correct sam-ba binary in flasher.sh script
Instead of using the install of sam-ba under host/opt directly, use the symlink
created in host/bin. The side effect of doing this instead allows the correct
sam-ba binary to be used based on the host arch being 32 bit or 64 bit.
Signed-off-by: Joshua Henderson <joshua.henderson@microchip.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Bump sama5d3xek to linux4sam_5.8. The 3 foundation components have their tags
changed (AT91Bootstrap, U-Boot, Linux kernel). This requires bumping the kernel
header version and u-boot dtc as well.
at91bootstrap 3.8.10 is required to support gcc7.
Signed-off-by: Joshua Henderson <joshua.henderson@microchip.com> Cc: Nicolas Ferre <nicolas.ferre@microchip.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
mesa3d: properly propagate the dependencies of BR2_PACKAGE_LLVM
We cannot simply select BR2_PACKAGE_LLVM, we need to take into account
all its dependencies. BR2_PACKAGE_LLVM_ARCH_SUPPORTS was already taken
into account, but not the other dependencies.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com> Acked-by: Valentin Korenblit <valentin.korenblit@smile.fr> Signed-off-by: Thomas Petazzoni <a class="moz-txt-link-rfc2396E" href="mailto:thomas.petazzoni@bootlin.com"><thomas.petazzoni@bootlin.com></a></pre> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
llvm: add dependency on !BR2_TOOLCHAIN_HAS_GCC_BUG_64735
LLVM uses std::shared_future, which until gcc 7.x isn't available on
architectures that don't provide lock-free atomics, see
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64735.
Buildroot already has a BR2_TOOLCHAIN_HAS_GCC_BUG_64735 option to
handle such a case, so simply add this new dependency to LLVM. It will
make sure LLVM doesn't get built on ARMv5 platforms using a gcc older
than 7.x.
When the internal PCRE library of PHP is used, it tries to use a JIT
engine, which is only available on some architectures.
However, the mechanism used to disable JIT has changed in recent PHP
versions, and it now has a proper --without-pcre-jit option. Switch
over to that to properly disable JIT on unsupported platforms.
It has been tested to fix the build of PHP on ARC and Microblaze.
Since c8ef0c03b0b (download: put most of the infra in dl-wrapper), the
backend for local files is now named after the scheme, which is 'file'
for a local file.
>From the same commit on, the directory part and the basename are now
passed separately, to let the backend reconstruct the full path when it
needs to do so, which is the case for the 'file' backend too.
Finaly, ff559846fdc1 (support/download: Add support to pass options
directly to downloaders) introduced a nasty error, as it made use of
"${@}" when calling its internal function. Revert that mess now...
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr> Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com> Cc: Peter Korsgaard <peter@korsgaard.com> Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com> Cc: Arnout Vandecappelle <arnout@mind.be> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
PIE breaks the build on m68k and Microblaze, so let's add a patch that
disables using PIE by default in keepalived. Obviously, when we start
seriously testing the BR2_RELRO_FULL option, we will realize that it
is broken on some architectures, and will probably disable it on m68k
and Microblaze.
We can drop 0001-Fix-build-with-Linux-kernel-headers-v4.15.patch as it
has been merged upstream, as of commit
https://github.com/acassen/keepalived/commit/a5457d3a1f579fef5825ffe841f1613b0014ff19.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Adam Duskett [Mon, 12 Mar 2018 10:12:31 +0000 (06:12 -0400)]
python-psycopg2: new package
Psycopg is the most popular PostgreSQL database adapter for the Python
programming language. Its main features are the complete implementation of the
Python DB API 2.0 specification and the thread safety (several threads can
share the same connection). It was designed for heavily multi-threaded
applications that create and destroy lots of cursors and make a large number of
concurrent INSERTs or UPDATEs.
Signed-off-by: Adam Duskett <aduskett@gmail.com> Reviewed-by: Yegor Yefremov <yegorslists@googlemail.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
postgresql: propagate BR2_USE_MMU dependency to Config.in comment
The BR2_PACKAGE_POSTGRESQL option depends on BR2_USE_MMU, so the
Config.in comment about the dynamic library dependency should only be
displayed if the BR2_USE_MMU requirement is met.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
$(dir does not strip last path separator and the wget wrapper adds a '/'
between the -u argument and the filename when it reconstructs the URL,
leading us to have two slashes before the filename when accessing the
upstream location.
Most upstream servers ignores this, but not all - Leading to download issues
from pypi.
Notice: As pointed out by Arnout, we cannot simply use $(PKG)_SITE here as
_PATCH and _EXTRA_DOWNLOADS may contain full URLs.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
download/git: fix basename for files inside tarballs
Commit "6d938bcb52 download: git: introduce cache feature" introduced a
typo that makes the tarball to contain files without the package
basename:
$ tar -tvf good-a238b1dfcd825d47d834af3c5223417c8411d90d.tar.gz
-rw-r--r-- 0/0 8 2017-10-14 02:10 ./file
The hashes in the tree were calculated with the basename.
In the most common scenario, after the download ends the tarball is
generated, the hash mismatches and the download mechanism falls back to
use the tarball from http://sources.buildroot.net .
The problem can be reproduced by forcing the download of any git package
PKG that has a hash file to check against:
$ make defconfig
$ ./utils/config --set-str BR2_BACKUP_SITE ""
$ BR2_DL_DIR=$(mktemp -d) make PKG-dirclean PKG-source
Fix the typo so the basename is really added to the files, that was
clearly the intention of the code.
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com> Cc: Arnout Vandecappelle <arnout@mind.be> Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com> Cc: Peter Korsgaard <peter@korsgaard.com> Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com> Cc: Yann E. MORIN <yann.morin.1998@free.fr> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Giulio Benetti [Mon, 13 Nov 2017 18:41:55 +0000 (19:41 +0100)]
qt5base: add support for sunxi-mali{,-mainline} eglfs backend
Only one of the two can be selected (they are conflicting virtual
packages) so testing against a single 'y' is enough.
Signed-off-by: Giulio Benetti <giulio.benetti@micronovasrl.com>
[Arnout: add sunxi-mali and sunxi-mali-mainline in a single patch] Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Giulio Benetti [Mon, 13 Nov 2017 18:41:54 +0000 (19:41 +0100)]
qt5base: correct eglfs support in qmake.conf.in
Setting EGLFS_DEVICE_INTEGRATION at the end of qmake.conf like is done
by commit 0c219ddb8a doesn't work correctly: it has to be set before the
include(../common/linux_device_post.conf)
Instead of appending to the file, change it into a qmake.conf.in
template file that contains a placeholder for the
EGLFS_DEVICE_INTEGRATION assignment and update it with sed. Since the
sed always has to be executed, this removes the need for a separate
QT5BASE_CONFIGURE_QMAKE_CONFIG definition.
Signed-off-by: Giulio Benetti <giulio.benetti@micronovasrl.com>
[Arnout: simplify the replacement, move sunxi-mali support to a
separate patch] Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Adam Duskett [Mon, 12 Mar 2018 10:12:30 +0000 (06:12 -0400)]
postgresql: add version output to pg_config
Some external packages call pg_config to determine the installed PostgreSQL
version. Add this output to Buildroots own pg_config, so these packages
correctly compile.
Signed-off-by: Adam Duskett <aduskett@gmail.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Peter Korsgaard [Wed, 4 Apr 2018 15:51:32 +0000 (17:51 +0200)]
python-webpy: security bump to version 0.39
>From the changelog:
2018-02-28 0.39
* Fixed a security issue with the form module (tx Orange Tsai)
* Fixed a security issue with the db module (tx Adrián Brav and Orange Tsai)
2016-07-08 0.38
..
* Fixed a potential remote exeution risk in `reparam` (tx Adrián Brav)
License files are still not included on pypi, so continue to use the git
repo. Upstream has unfortunately not tagged 0.39, so use the latest commit
on the 0.39 branch. A request to fix this has been submitted:
https://github.com/webpy/webpy/issues/449
0.39 now uses setuptools, so change the _SETUP_TYPE.
Add hashes for the license files.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
This patch provides LLVM support for Mesa3D, enabling llvmpipe
software rasterizer if Gallium swrast is selected.
In case Gallium r600 is selected, llvm AMDGPU backend is built.
Having llvm installed also enables radeonsi Gallium driver, but
it is not provided with this patch as it hasn't been tested.
It uses llvm-config (host variant) installed in STAGING_DIR/usr/bin
to get LLVM libraries. Assuming that LLVM version 5.0.1 is installed,
llvm-config --libs will output -lLLVM-5.0.
This patch installs LLVM tools and libraries for the host and
libLLVM.so for the target.
In order to cross-compile LLVM for the target, LLVM
must be installed on the host, or at least llvm-tblgen.
This is necessary as the path to host's llvm-tblgen must
be specified when cross-compiling using the LLVM_TABLEGEN option.
Also, a version of llvm-config that can run on the host will
be required by packages that link with LLVM libraries, so we
need to generate it and install it in STAGING_DIR/usr/bin.
It is important to remark why we need llvm-config(host variant)
installed in STAGING dir. This tool is necessary to build
applications that use LLVM, as it prints the compiler flags,
linker flags and object libraries needed to link against LLVM libs.
More info: https://bugs.chromium.org/p/chromium/issues/detail?id=219369
The original idea was to compile only llvm-tblgen and llvm-config
for the host, as they are the only necessary components. However,
llvm-config tool does not work as expected if it is not linked with
libLLVM.so, so we must also enable LLVM_LINK_LLVM_DYLIB, what builds
LLVM as a single shared library and links LLVM tools with it.
More info: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=224847
in comment #11.
If we don't build full LLVM for the host, it would be necessary to
patch configure.ac from mesa3d if we want dynamic linking, because it
uses llvm-config (host variant installed in STAGING_DIR) to get the
necessary LLVM libraries to link with, which has the following problems:
- llvm-config --shared mode outputs static (even if LLVM is built as
one shared library) which leads to link issues with libgallium.
- llvm-config --libs outputs all LLVM tiny libs: -lLLVMLTO,
-lLLVMPasses,etc instead of the single shared library containing
all LLVM components (-lLLVM-5.0)
Mesa tries to execute: llvm-config --link-shared --libs, but this outputs
llvm-config: error: libLLVM-5.0.so is missing.
Given that these problems may arise with other packages that use LLVM,
it is preferable to do a full build for the host. Also, having a
complete installation of LLVM on the host will also facilitate the
integration of Clang front-end, which is going to be added in a future
patch.
As option LLVM_BUILD_LLVM_DYLIB is enabled for the llvm target variant,
a single shared library containing all LLVM components is built.
This option is not compatible with BUILD_SHARED_LIBS, which generates
one .so per library and is only recommended for use by LLVM developers.
Tools and utils are not built for the target. The patch aims to provide
LLVM support for other packages.
The main options needed to cross-compile LLVM are the following ones:
Signed-off-by: Valentin Korenblit <valentin.korenblit@smile.fr>
[Thomas:
- add dependency on thread and C++ and update the Config.in comment
accordingly.
- make the Config.in comment depend on BR2_PACKAGE_LLVM_ARCH_SUPPORTS
so that it isn't disabled on architectures where LLVM is anyway not
supported.] Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Thomas Petazzoni [Fri, 23 Mar 2018 20:54:53 +0000 (21:54 +0100)]
support/scripts/pkg-stats-new: add current version information
This commit adds a new column in the HTML output containing the
current version of a package in Buildroot. As such, it isn't terribly
useful, but combined with the latest upstream version added in a
follow-up commit, it will become very useful.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com> Reviewed-by: Ricardo Martincoski <ricardo.martincoski@gmail.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Thomas Petazzoni [Fri, 23 Mar 2018 20:54:52 +0000 (21:54 +0100)]
support/scripts/pkg-stats-new: add -n and -p options
This commit adds the following options to the pkg-stats-new script:
-n, to specify a number of packages to parse instead of all packages
-p, to specify a list of packages (comma-separated) to parse instead
of all packages
These options are basically only useful when debugging/developing
this script, but they are very useful, because the script is rather
slow to run completely with all 2000+ packages, especially once
upstream versions will be fetched from release-monitoring.org.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com> Reviewed-by: Ricardo Martincoski <ricardo.martincoski@gmail.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Thomas Petazzoni [Fri, 23 Mar 2018 20:54:51 +0000 (21:54 +0100)]
support/scripts/pkg-stats-new: rewrite in Python
This commit adds a new version of the pkg-stats script, rewritten in
Python. It is for now implemented in a separate file called,
pkg-stats-new, in order to make the diff easily readable. A future
commit will rename it to pkg-stats.
Compared to the existing shell-based pkg-stats script, the
functionality and output is basically the same. The main difference is
that the output no longer goes to stdout, but to the file passed as
argument using the -o option. This allows stdout to be used for more
debugging related information.
The way the script works is that a first function get_pkglist()
returns a list of Package objects. Then, the function
package_init_make_info() uses 'make printvars' to gather information
about all packages, stored as class variables in the Package
class. Then, we iterate over all packages, and use various methods of
the Package class to retrieve all details about the package:
infrastructure, presence of hash file, presence of license
information, etc.
calculate_stats() then calculates global statistics (how packages have
license information, how packages have a hash file, etc.). Finally,
dump_html() produces the HTML output, using a number of sub-functions.
One improvement over the shell-based version is that we can use
regexps to exclude some .mk files. Thanks to this, we can exclude all
linux-ext-*.mk files, avoiding incorrect matches.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com> Reviewed-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
projects / coffee / buildroot.git / log