"The Graphite framework for loop optimizations no longer requires the
CLooG library, only ISL version 0.14 (recommended) or 0.12.2. The
installation manual contains more information about requirements to
build GCC."
This change helps to avoid version badness.
Signed-off-by: Bryan Hundven <bryanhundven@gmail.com>
This change needs another change from pull request #81, but it's kind of
a chicken/egg scenario. The 'select's in CC_GCC_5_1 need to be
refactored a bit, and would be easier to test if gcc-5.1.0 was commited.
Most of the refactoring will happen with CC_GCC_HAS_GRAPHITE.
Signed-off-by: Bryan Hundven <bryanhundven@gmail.com>
In commit cd47c091ba
I had forgot to also remove the config/libc/eglibc.in.
This commit removes it.
Signed-off-by: Bryan Hundven <bryanhundven@gmail.com>
There are other languages which work with bare metal compilers.
As an example crosstool-ng is recommended to build D/GDC bare metal
compilers.
Signed-off-by: Johannes Pfau <johannespfau@gmail.com>
This commit updates to the latest longterm and stable kernel versions as
of March 15, 2015.
Signed-off-by: Cristoforo Cataldo <cristoforo.cataldo@gmail.com>
This commit updates to the latest longterm and stable kernel versions as
of February 18, 2015.
Signed-off-by: Cristoforo Cataldo <cristoforo.cataldo@gmail.com>
As posted on http://www.eglibc.org/
====================
EGLIBC is no longer developed and such goals are now being addressed
directly in GLIBC.
====================
I'm not interested in maintaining build support for unsupported
software.
Older branches of crosstool-ng continue to have eglibc support.
If you find issues with older branches, I'm always open to pull
requests.
Removing eglibc also frees up glibc cleanup and build optimization.
Signed-off-by: Bryan Hundven <bryanhundven@gmail.com>
This commit updates to the latest longterm and stable kernel versions as
of January 16, 2015 and adds also 3.18 version.
Signed-off-by: Cristoforo Cataldo <cristoforo.cataldo@gmail.com>
We had following problem: We're building a toolchain with an old glibc
version for compatibility with old Linux distributions (glibc 2.9). This
version requires make < 4 to build. However, the configure script of
glibc looks for make in the order "gnumake", "gmake" and "make". So when
"gmake" is available in the system (which is the case on Gentoo Linux
per default, unfortunately), then configure finds the system gmake 4.1
instead of the ct-ng make 3.82.
This patch adds an option to install a symlink so that 'gmake' is also
available in the old version when building toolchains.
Signed-off-by: Bernhard Walle <bernhard@bwalle.de>
Can safely skip the core pass-1 for normal baremetal builds,
but when building a canadian baremetal, the repair_cc
functionality (GCC_FOR_TARGET) in gcc.sh will force the
core pass-2 to attempt to build gcc and libgcc without a
${CT_TARGET}-gcc existing, causing a failure on
${CT_TARGET}-gcc -dumpspecs > tmp-specs
Signed-off-by: David Holsgrove <david.holsgrove@xilinx.com>
This commit updates to the latest longterm and stable kernel versions as of
December 10, 2014.
Signed-off-by: James Byrne <james.byrne@origamienergy.com>
As per the glibc release notes for 2.20:
--------------------------------------------------------------------------------
All supported architectures now use the main glibc sysdeps directory
instead of some being in a separate "ports" directory (which was
distributed separately before glibc 2.17).
--------------------------------------------------------------------------------
There isn't a ports directory anymore. So disable using and forcing it.
closes#7 on crosstool-ng github
Signed-off-by: Bryan Hundven <bryanhundven@gmail.com>
I messed up the previous commit.
I only updated half the config file, and the latest 4.8 is 2014.08.
🤦
Signed-off-by: Bryan Hundven <bryanhundven@gmail.com>
CUSTOM_LOCATION config options only presented in menuconfig if component
CUSTOM version selected.
Signed-off-by: David Holsgrove <david.holsgrove@xilinx.com>
CUSTOM_LOCATION config options only presented in menuconfig if component
CUSTOM version selected.
Signed-off-by: David Holsgrove <david.holsgrove@xilinx.com>
No longer recommended practice to use --enable-add-ons=nptl, so
for 2.20 and later (along with custom glibc), don't add the
CT_THREADS to the addons_list
https://sourceware.org/glibc/wiki/Release/2.20#Packaging_Changes
Signed-off-by: David Holsgrove <david.holsgrove@xilinx.com>
This commit adds a configuration knob for enabling extra developer
warnings to be enabled during the musl-libc build.
Signed-off-by: Bryan Hundven <bryanhundven@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
This option enables a configuration knob for adding debugging info.
Signed-off-by: Bryan Hundven <bryanhundven@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
This change removes 1.0.3 and 1.1.3 and linker regession patches for
those versions.
We add 1.0.4, and a patch needed for gcc-4.9.x which defines
`max_align_t'.
Signed-off-by: Bryan Hundven <bryanhundven@gmail.com>
libsaniotizer requires a few headers that are not in uClibc, for
example. Also, it is only available for native threads (NPTL under
glibc.) Finally, it is only available starting with gcc-4.8.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
This patch adds initial support for musl-libc.
Musl-libc versions currently supported:
* 1.0.3 (Stable)
* 1.1.3 (Previous Mainline)
* 1.1.4 (Mainline)
Futher improvements are needed.
* gcc-4.9.x has issues (Might be fixed in musl-1.1.4).
* Multilib support is needed.
* Checks to make sure paths are correct.
* Move to 2-step gcc build. 3-step build is not necessary.
Signed-off-by: Bryan Hundven <bryanhundven@gmail.com>
[yann.morin.1998@free.fr: removed the gcc musl patch, to be added later;
removed dead code do_get_arch()]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
There is no need to differentiate the win32 threads case, since we
can cosider them to be the native implementation on Windows.
Besides, with the previous patch, nothing uses it anymore.
So, just remove it.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc Bryan Hundven <bryanhundven@gmail.com>
This will help add new implementations, such as the one in musl.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Bryan Hundven <bryanhundven@gmail.com>
Use a more coherent naming for the options. This will help commonalise
the native case (e.g. NPTL on Linux, win32 on Windows), and add alternate
implementations (e.g. musl.)
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Bryan Hundven <bryanhundven@gmail.com>
This change adds support for experimental patches to be introduced to
crosstool-ng. The patches enabled by this option are to be located here:
patches/experimental/<package>/<version>/XXXX-NAME.patch
Where, XXXX is the patch number to be applied in order, like:
0001-some_patch_one.patch
0002-some_patch_two.patch
9999-some_patch_to_be_applied_last.patch
In the first patch series, all patches in the EXPERIMENTAL_PATCHES
option will be applied all at once, or none at all.
In a later [RFC] patch, I plan on adding finer tuned patch
enable/disable options based on the name of the patch and where it is
located in the patches/experimental sub-tree. So the name of the patch
should use underscores between words in the patch name.
Signed-off-by: Bryan Hundven <bryanhundven@gmail.com>
[yann.morin.1998@free.fr: slightly reword prompt]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
In make-3.8x, the $(wildacrd) function would sort the entries,
while in make-4.x, it would just return the entries in any
unpredictable order [*]
Use the $(sort) function to get reproducible behaviour.
[*] Well, most probably the roder the entries appear when read
from readdir()
Reported-by: Andrew Ruder <andrew.ruder@elecsyscorp.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Tested-by: Andrew Ruder <andrew.ruder@elecsyscorp.com>
Those versions are no longer available upstream. They have purely and
simply disapeared, without leaving any trace of their mere existences.
Just keep the latest cloog-ppl-0.15.11, which still exists on the gcc
infra mirror (but for how long?)
Reported-by: Guillaume FLORENCE-COURTAND <gflorenc@laposte.net>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Technically, I don't forbid powerpcle support either, but I'm not sure that
there is any library/compiler support for that at the moment (though the hw
technically makes it possible).
powerpc64le needs glibc 2.19 and gcc 4.9. I haven't looked into the support
tools, but at least gdb 7.5 is too old (7.7.1 definitely has support).
Also make powerpc64 non-experimental. It's practically old at this point.
Signed-off-by: Cody P Schafer <dev@codyps.com>
[yann.morin.1998@free.fr: use ${target_endian_le} and ${target_bits_64}]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Message-Id: <64bfbbced9dd8f62e0d6.1399801945@gun>
Patchwork-Id: 347775
Signed-off-by: Cody P Schafer <dev@codyps.com>
[yann.morin.1998@free.fr: latest is now a 4.9]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Message-Id: <5bac788539bb272893ed.1399801933@gun>
Patchwork-Id: 347774
Some of the compiler options depend on the C library choice, (e.g.
whether core passes are needed).
Since the compiler menu comes before the C library menu, those options
may or may not be visible until the C library is chosen, leading to
either options being visible by the user (thus be puzzling as they would
not apply to his case), or invisible to him (yet again puzzling him).
Invert the order of the compiler and the C library in the menuconfig. It
anyway looks more rational, in the end.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
The final bare-metal compiler is built using the core backend.
Currently the core uses the CC_CORE_EXTRA_CONFIG_ARRAY variable.
While this works as supposed to, this can leave the user puzzled
in the menuconfig, since all he can see is the core options, not
the final options.
Only show the core options if any of the core passes are needed,
and use the final options in the core-backend if we're issuing
the bare-metal compiler.
Signed-off-by: Cody P Schafer <dev@codyps.com>
[yann.morin.1998@free.fr: hide core options if no core pass needed;
use final option in core backend if issuing the bare-metal compiler]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Message-Id: <22181e546ba746202489.1399688067@localhost>
Patchwork-Id: 347586
Add an ARCH_FLOAT_AUTO option to avoid passing float options
to gcc and allow it to choose
Signed-off-by: Cody P Schafer <dev@codyps.com>
[yann.morin.1998@free.fr: slightly reword the commit log]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Message-Id: <088489641f4790262685.1399687813@localhost>
Patchwork-Id: 347584
Allow '-1' to be specified as CONNECTION_TIMEOUT to disable the use
of the connection timeout for wget.
Signed-off-by: Cody P Schafer <dev@codyps.com>
Message-Id: <cb33f8c2cbaf802d4f04.1399687632@localhost>
Patchwork-Id: 347582
update linux kernel headers to a newer version
Signed-off-by: '"Daniel Zimmermann" <netzimme@gmail.com>'
[yann.morin.1998@free.fr: further bump the versions]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Message-Id: <424a44a11e7051d8d894.1392963455@haus-VirtualBox>
Patchwork-Id: 322417
In some cases, building the core pass-1 is unneeded, as the C library
will not try to build anything when installing headers, or sill not
install anything at all.
This is for example the case for newlib, wihch does not require a core
pass-1 since it builds nothing and installs no header.
This should also be the case for newer glibc-es with newer gcc-es,
which no longer require a core pass-1, since the circular dependency
glibc <-> gcc (about TLS?) has been resolved.
Reported-by: Trevor Woerner <trevor.woerner@linaro.org>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
On systems with make-3.82, we forcibly force the build and the use
of make-3.81
But some newer tools break when building with make-3.81. For example,
eglibc-3.18 breaks.
Introduce a new blind options that tools may select if they require
make-3.81. If the system does not have make-3.81, and this option is
selected, then we force the build of make-3.81. Otherwise, we leave
it to the user to decide on his own.
Note that no component selects this option for now. It will come in
later patches as we find them.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Ray Donnelly <mingw.android@gmail.com>
[yann.morin.1998@free.fr: add message "checking for make 3.81"]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Message-Id: <8b8bf6998f3d239f6c74.1392920971@advancedsearch.virginmedia.com>
Patchwork-Id: 322302
pump ltrace to version 0.7.3
backport a patch from Peter Wu to silence warnings/errors in ltrace
Signed-off-by: '"Peter Wu" <lekensteyn@gmail.com>'
Signed-off-by: '"Daniel Zimmermann" <netzimme@gmail.com>'
[yann.morin.1998@free.fr: remove superfluous empty lines]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Message-Id: <8349c35b506aa79c4e69.1392845765@haus-VirtualBox>
Patchwork-Id: 322017
Decimal floats need support form the C library, namely support
for fenv, which is missing in uClibc for any architecture but
x86/32.
Add an option (a choice) to enable or disable decimal floats.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
LinuxThreads are dead. Buried 6 feet under. Long forgotten.
Time for mourning has to come to an end.
But fear not, younster, for we have a great successor as NPTL!
Times for rejoicing have come, now!
LinuxThreads are dead. Long live NPTL!
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Support for fenv.h is a little bit more tricky that enabling it only
for x86-32 is not right.
Add an option for the user to choose whther to install fenv.h or not.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
No need to 'select' gold and plugins in versions later
than 2.21, since we inherit them with 2_21_or_later.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
AArch64 id the 64-bit variant for ARM.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Zhenqiang Chen <zhenqiang.chen@linaro.org>
Cc: Michael Hope <michael.hope@linaro.org>
Thus spake Niels:
This commit adds support for binutils 2.23.2
http://crosstool-ng.org/hg/crosstool-ng/rev/bc4b36e3df0c
However, in 'config/binutils/binutils.in <http://binutils.in>' the
Kconfig option to select this version of binutils does not automatically
select the symbol 'BINUTILS_2_23_or_later'. This makes gold unavailable
for the latest version of binutils. I'm not sure whether this is on
purpose or not, my impression is that you forgot to add the line "select
BINUTILS_2_23_or_later".
The problem still persists in the latest version in the repository
(032ebcc89c96).
Reported-by: Niels Penneman <niels@penneman.org>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Pump version for strace to 4.7 and 4.8.
Add patch from strace mainline.
Work around conflict between <sys/ptrace.h> and <linux/ptrace.h>
Since glibc-2.18~39 <sys/ptrace.h> defines ptrace_peeksiginfo_args
which collides with <linux/ptrace.h>.
* configure.ac: Check for `struct ptrace_peeksiginfo_args' in
<sys/ptrace.h>.
* process.c: Work around potential conflict between <sys/ptrace.h>
and <linux/ptrace.h> by redefining ptrace_peeksiginfo_args.
* signal.c: Likewise.
* syscall.c: Likewise.
* util.c: Likewise.
0b4060f61f/
Signed-off-by: "Daniel Zimmermann" <netzimme@gmail.com>
Message-Id: <67b082cf1cdc8276eb4a.1385366288@haus-VirtualBox>
Patchwork-Id: 293842
Don't download glibc-ports when glibc or eglibc version greater than 2.16,
because the "ports" source is mainline in the glibc or eglibc since version 2.17.
Signed-off-by: "Daniel Zimmermann" <netzimme@gmail.com>
Message-Id: <9c045ca3cf1b9dc89da3.1384602843@haus-VirtualBox>
Patchwork-Id: 291766
[yann.morin.1998@free.fr: slightly tweak subject, change variable name]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cset #15eedf548d33 (cc/gcc: add preliminray support for 4.8) left a
lingering 'select CLOOG_NEEDED', which should have been removed.
Fix that.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Update Linaro GCC with the latest available revisions.
Also updates Linaro GCC 4.6 to the final version.
Signed-off-by: "Michael Hope" <michaelh@juju.net.nz>
Message-Id: <83e33862f7641c2ad326.1372451435@localhost>
Patchwork-Id: 255666
CLooG 0.18+ will use ISL instead of PPL, so we have to configure
adequately depending of which backend is in use.
The Kconfig entries will decide for us which is selected, so we
can rely on either PPL xor ISL to be selected, not both.
Reported-by: "Plotnikov Dmitry" <leitz@ispras.ru>
[Dmitry did a preliminray patch to add ISL support,
which this patch is inspired from]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
This means:
- introduce the new symbols for 4.8
- do not always select PPL if graphite is selected
Reported-by: "Plotnikov Dmitry" <leitz@ispras.ru>
[Dmitry did a preliminray patch to add gcc-4.8 support,
which this patch is inspired from]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
ISL is used by gcc-4.8 onward for GRAPHITE, so is also used as
backend for CLooG 0.18.0 onward.
Reported-by: "Plotnikov Dmitry" <leitz@ispras.ru>
[Dmitry did a preliminray patch to add ISL, which this one is inspired from]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Update Linux with the latest available revisions.
Signed-off-by: "Benoît Thébaudeau" <benoit.thebaudeau@advansee.com>
[yann.morin.1998@free.fr: added newer versions released since]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Message-Id: <435dcc4cade342ab9ee0.1365686746@advdt005-ubuntu>
Patchwork-Id: 235741
Update Linux with the latest available revisions.
Signed-off-by: "Benoît Thébaudeau" <benoit.thebaudeau@advansee.com>
Message-Id: <14c04210a1dc18f3c678.1363295061@advdt005-ubuntu>
Patchwork-Id: 227803
Update Linux with the latest available revisions.
Signed-off-by: "Benoît Thébaudeau" <benoit.thebaudeau@advansee.com>
[yann.morin.1998@free.fr: add latest versions since released]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Message-Id: <125b3612bbefcb57166b.1361275815@advdt005-ubuntu>
Patchwork-Id: 221686
Now we use defconfig files to store the samples, we have to be a bit more
conservatives in the symbols names, so as to avoid gigantic version bumps
when updating sub-level versions from a package.
Update samples accordingly.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Now we use defconfig files to store the samples, we have to be a bit more
conservatives in the symbols names, so as to avoid gigantic version bumps
when updating sub-level versions from a package.
For example (using fictitious versions):
- in crosstool-NG 1.17.0, we choose:
- latest gcc is gcc-linaro-4.7-2012.10, which is the default for the
choice in the menuconfig
- gcc-linaro-4.6-2012.10 is selected
- so, sample has an explicit symbol for the selected gcc version, as it
is not the default
- we update to crosstool-NG 1.18.0:
- latest gcc version is gcc-linaro-4.7-2013.01
- gcc-linaro-46 has been updated to gcc-linaro-4.6-2013.01
- as the sample now has no *valid* symbol to set the gcc version, the
default is used, while we would have expected to still use the 4.6
release from linaro, not the 4.7
Get rid of sub-level (ie. the third digit sequence in versions) from the
symbols for linaro versions.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
For some architectures, it is legit to have an alternate value in the
'architecture' part of the tuple. For example:
armv5te-*
armv7a8-*
Besides, some packages expect the tuple to reflect the arch variant
(eg. openMPI) to detect the variant's capabilities (eg. atomic
primitives).
This patch adds an option for the user to specify a suffix to be added
to the arch-part of the tuple.
Signed-off-by: Willy Tarreau <w@1wt.eu>
Message-ID: <20130120225822.GS6838@1wt.eu>
Patch-Id: 213994
[yann.morin.1998@free.fr: make it a suffix, not an override]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
kernel/linux: update revisions
Update Linux with the latest available revisions.
Signed-off-by: "Benoît Thébaudeau" <benoit.thebaudeau@advansee.com>
Message-Id: <df032717ca91dc9cc876.1358518690@advdt005-ubuntu>
Patchwork-Id: 213616
Running as root is really, really dangerous.
Add a runtime-check that refuses to build if running as root.
Can be overriden with a double switch in the menuconfig.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Well, leave the prompt as an OBSOLETE thing, scheduled to
be removed soon.
As an indication OABI lives its last days, gcc-4.8 will no
longer recognise non-EABI targets.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Toolchains that use the hard-float ABI now are to be denoted by a tuple
ending in *eabihf, while the prevbious *eabi is now an indication that
the toolchain uses the softfloat ABI.
This is purely a cosmetic thing, for distros to differentiate their
hardfloat-ABI ports from their softfloat-ABI ports.
(note: softfloat ABI does not mean that it is using softfloats; it can
be using hardfloat instructions, but using the softfloat ABI).
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
