Make the gcc build system create multiple libgcc.a instances for
different ABIs.
Change-Id: I1c888bf751bf43566da8927ed0aedb53857363bf
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Reviewed-on: https://review.coreboot.org/13625
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Not much testing, update mostly so we can test with the
latest scan-build.
Change-Id: I50d28b7e0dfd31f9ae565c8515d5ab1760ca4c62
Signed-off-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-on: https://review.coreboot.org/13516
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
The compiler really supports a whole line of ARM CPUs, not just
ARMv7a:
arm-eabi-gcc: note: valid arguments to '-march=' are: armv2 armv2a
armv3 armv3m armv4 armv4t armv5 armv5e armv5t armv5te armv6 armv6-m
armv6j armv6k armv6s-m armv6t2 armv6z armv6zk armv7 armv7-a armv7-m
armv7-r armv7e-m armv7ve armv8-a armv8-a+crc iwmmxt iwmmxt2 native
So let's reflect this in the cross compiler name.
Change-Id: I717760d80954655b2de9ae019b813d81e9a75762
Signed-off-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-on: https://review.coreboot.org/13515
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
It may still fail on non-Linux, and the compiler may do fancy things,
but it builds.
Change-Id: If3456f5fef8d01082a49978dc7cda5450f96f5cc
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Reviewed-on: https://review.coreboot.org/13416
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Reviewed-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
GMP's configure tries to build for 64-bit with a 32-bit userspace on
NetBSD too. Help it by forcing ABI=32.
Change-Id: I290ea0ef1626fdd88dc3ff74fadb9578ef6a1c9c
Signed-off-by: Nico Huber <nico.h@gmx.de>
Reviewed-on: https://review.coreboot.org/13067
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
This reverts commit 89798bcb0c.
Disable building gnat again as it turned out that many distros don't
ship with a sufficient recent version of gnat. We'll have to find a
reliable way to check for the installed gnat version and query the
user or bootstrap gcc in that case.
Change-Id: Ife7cf7c9d1567aca898ce308b120a7b9e146e5f5
Signed-off-by: Nico Huber <nico.huber@secunet.com>
Reviewed-on: https://review.coreboot.org/13422
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-by: Martin Roth <martinroth@google.com>
Reviewed-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Tested-by: build bot (Jenkins)
I thought we'd be using gnatmake but it's deprecated. Who needs it
anyway?
Change-Id: Ic08add72e771fa346c8a736ea901863ea5737d91
Signed-off-by: Nico Huber <nico.h@gmx.de>
Reviewed-on: https://review.coreboot.org/13041
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Instead of printing out a single tool that needs to be installed
each time buildgcc is run, print out the entire list of tools
to be installed, then halt.
Change-Id: I7761760eef3c45ba371f882a4f987408945bb3e5
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/12856
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Previously, when we tested for g++ and two different versions of clang,
if the earlier versions were not found, buildgcc would still request
that they be installed. This obviously isn't needed, and isn't the
desired outcome.
Now, if one of the first tests fails, nothing gets printed. If all
the tests fail, it tells you to install either g++ or clang.
Change-Id: I71359f59c4c6bee3c3c55e4e6105f11e6ca51527
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/12852
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
This reverts commit 68d0e4a5a1.
Special handling of MPFR is no longer needed with the latest
MPFR release.
Change-Id: I96d9ea92cfb74eed6af2ba62254f0678081e2b4f
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: https://review.coreboot.org/12833
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
The current MPFR version contains a stale config.guess file
that requires special handling on ppc64el systems. Bump
the MPFR version to the latest release.
Change-Id: I5e86c732c09f8a6a43f9812452124d64d337ea3f
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: https://review.coreboot.org/12832
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
The config.guess file included with MPFR is completely obsolete,
leading to build failures on ppc64el due to the system architecture
not being detected. Regenerate the files from the host system via
automake before attempting to build MPFR.
Change-Id: I00fc16003906e373d112c25978197ac907adccfd
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: https://review.coreboot.org/12816
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
The previous official GMP release (6.0.0) contains a bug that
prevents compilation on ppc64el systems. Increase version
to the latest version (6.1.0).
Bug details:
gcc build on ppc64el fails with:
(.text+0x4c): undefined reference to `BMOD_1_TO_MOD_1_THRESHOLD'
While I don't have an exact commit hash due to Hg use upstream,
a missing BMOD_1_TO_MOD_1_THRESHOLD define on ppc64el was quietly
fixed in Hg before the 6.1.0 release.
Change-Id: I1c05a1c194141db5f8522148c2e20e7558d34714
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: https://review.coreboot.org/12811
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Martin Roth <martinroth@google.com>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
clang version now returns:
coreboot toolchain v1.33 November 25th, 2015 clang version 3.6.1
(tags/RELEASE_361/final) (based on LLVM 3.6.1)
Change-Id: I948d7f4d06c244987342cfc7d5c7e728cbed93bd
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/12777
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Add the coreboot toolchain version to iasl's version output.
% ./xgcc/bin/iasl -v
Intel ACPI Component Architecture
ASL+ Optimizing Compiler version 20150619-64
Copyright (c) 2000 - 2015 Intel Corporation
coreboot toolchain v1.33 November 25th, 2015
This won't actually be checked until the next version of
iasl so that we don't have to rebuild again for no reason.
The buildgcc version was intentionally not incremented for
this minor change.
Change-Id: I03a1a777fdb84e34bfceb7b1eb43fffbc1f3a2fc
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/12688
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
By adding the version number of tools, we can help people keep up to
date with their tool versions. This will be used now to determine
whether the IASL version being used is the version supported by
coreboot.
Change-Id: I24a68b01c819871f90403869570125e71b96bd70
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/12545
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
It encourages users from writing to the FSF without giving an address.
Linux also prefers to drop that and their checkpatch.pl (that we
imported) looks out for that.
This is the result of util/scripts/no-fsf-addresses.sh with no further
editing.
Change-Id: Ie96faea295fe001911d77dbc51e9a6789558fbd6
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Reviewed-on: http://review.coreboot.org/11888
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Grep the output of wget, showing only the percentage.
Leave the final "100%" unerased.
Checking return code of wget is removed.
Change-Id: I4559e88d541738a594dce92e23589992f234cb9b
Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Signed-off-by: Zheng Bao <fishbaozi@gmail.com>
Reviewed-on: http://review.coreboot.org/11520
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
As with most other languages, a pre-installed Ada toolchain is needed
to build gcc's Ada frontend. To support building with older host tool-
chains, the patch `gcc-5.2.0_gnat.patch` disables warnings for unknown
pragmas. Building has been tested with host gcc-4.9 and hopefully works
with newer versions, too.
For convenience, the gnattools (e.g. gnatmake etc.) will be build if
'ada' is specified as a target language.
Change-Id: Ia78c29d1aba2943de5238421a324cfff8eb08875
Signed-off-by: Nico Huber <nico.huber@secunet.com>
Reviewed-on: http://review.coreboot.org/11590
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Add an option `--languages` which takes a list of target languages to
buildgcc. That list gets passed through to the configure step for
building gcc.
Also alter the Makefile to pass $(BUILD_LANGUAGES) to that option, if
this variable is set.
Change-Id: I6a74ab2c75871ea8d03a499cca33d88938b59c8d
Signed-off-by: Nico Huber <nico.huber@secunet.com>
Reviewed-on: http://review.coreboot.org/11589
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Change-Id: Ie3a44c6db9c9c186c52b4743334266ec5411ba8a
Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Signed-off-by: Zheng Bao <fishbaozi@gmail.com>
Reviewed-on: http://review.coreboot.org/11472
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Change-Id: I9f59a00e735f39df813b2216290da62eea3c595d
Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Signed-off-by: Zheng Bao <fishbaozi@gmail.com>
Reviewed-on: http://review.coreboot.org/11372
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Refactor the code to be better understandable.
Change-Id: Ia815a27f7cc83c226a32e87485d712a5fbf4168e
Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Signed-off-by: Zheng Bao <fishbaozi@gmail.com>
Reviewed-on: http://review.coreboot.org/11318
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
All compilers built, incl. x86_64-elf as multilib and riscv-elf.
Change-Id: Iafa61b1d2ffc9c737ab67a417c62417593b69372
Signed-off-by: Patrick Georgi <patrick@georgi-clan.de>
Reviewed-on: http://review.coreboot.org/10975
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
1. Add -P|--package to build iasl
2. Remove -G|--skip-gdb, which was to skip gdb.
3. Add -S|--scripting to build gdb
4. Remove -C|--clang, which was to build clang.
All these changes are aligned with the options parsing below.
The help text is correct.
Change-Id: I897ea5e8ab002086e45bf05ff33230815b246057
Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Signed-off-by: Zheng Bao <fishbaozi@gmail.com>
Reviewed-on: http://review.coreboot.org/11158
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Similar to what the below change says,
(
http://review.coreboot.org/10792
commit ddb8f80894
Author: Patrick Georgi <patrick@georgi-clan.de>
Date: Sat Jul 4 17:45:54 2015 +0200
buildgcc: Deal with gmp on 32bit Linux on 64bit CPUs
GMP is overeager to detect 64bit ABIs even if the entire running codebase is
32bit (but on a 64bit CPU). Enforce a 32bit build in that situation.
)
building GMP can not detect Cygwin is 32bit either if the
host which Cygwin is running is 64bit. We set ABI=32
in that case.
Change-Id: Ic53d75defebbe902325eb07f3d8631b2a53245ef
Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Signed-off-by: Zheng Bao <fishbaozi@gmail.com>
Reviewed-on: http://review.coreboot.org/11123
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Running `uname` on Cygwin gets "CYGWIN_NT-6.1-WOW" instead of "Cygwin".
We need to fix the $UNAME on Cygwin.
Change-Id: I540bfc52089951006fd0e20bb9893a3d891df9e1
Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Signed-off-by: Zheng Bao <fishbaozi@gmail.com>
Reviewed-on: http://review.coreboot.org/11124
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
It doesn't know "source", but wants the older "." instead
Change-Id: Iafa61b1d2ffc9c737ab67a417c62417593b69374
Signed-off-by: Patrick Georgi <patrick@georgi-clan.de>
Reviewed-on: http://review.coreboot.org/10974
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
In case of downloading errors, the URL is handy for analyzing the cause.
Change-Id: I6874cdc3c881cfdd52c80f80323536c30723654b
Signed-off-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-on: http://review.coreboot.org/10853
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <gaumless@gmail.com>
Running "clang -target i386-elf --print-librt-file-name" prints
[..]/bin/../lib/clang/3.6.1/lib/libclang_rt.builtins-i386.a
However, the correct path is [..]/lib/linux/libclang_rt.builtins-i386.a
on a Linux host. Hence, create symbolic links to make sure that our
build system finds the file where it expects it.
Change-Id: I21ef5c4a690d83c326717ca55c5ace558257a0ec
Signed-off-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-on: http://review.coreboot.org/10815
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
GMP is overeager to detect 64bit ABIs even if the entire running codebase is
32bit (but on a 64bit CPU). Enforce a 32bit build in that situation.
Change-Id: I23e9e57f3c8b0e3ad2e4e1e3eb106f7830aa76a1
Signed-off-by: Patrick Georgi <patrick@georgi-clan.de>
Reviewed-on: http://review.coreboot.org/10792
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
buildgcc fails if g++ or clang isn't found on the host. This
was failing on OSX due to the string used to check for clang
doesn't match "Apple LLVM". Add an additional search string for
clang "LLVM".
Change-Id: I05e36cfc690061b3233376d57f44f197cab933ea
Signed-off-by: Marc Jones <marc.jones@se-eng.com>
Reviewed-on: http://review.coreboot.org/10569
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
There's a separate target -P iasl for that now.
Change-Id: I95c0fe8fc266859d8a31b7bea890775dc9f19694
Signed-off-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-on: http://review.coreboot.org/10567
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
With this change, the x86_64-elf-gcc can compile i386-elf
binaries by specifying -m32. The patch against GCC is needed
to enable building the 32bit libraries when building x86_64-elf-gcc
Change-Id: Ic86a009eccfdf3e33a398bcdcc13b15c8dfc0d31
Signed-off-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-on: http://review.coreboot.org/10497
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
For consistency in user output, move the check for all
required utilities after printing the banner and parsing
options.
Change-Id: I5bf31368885c73e35f18b02d53d099f3f3871acc
Signed-off-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-on: http://review.coreboot.org/10566
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
When required tools are missing, try to give the user more detailed
information on how to solve the problem.
Change-Id: Ifa21c1af38a036a7d4f5a786041a87a7d45f4ec5
Signed-off-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-on: http://review.coreboot.org/10555
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
This moves the CMAKE definition down into the case statement
for $PACKAGE so that it is only required when the user wants to
build clang.
With this approach, "./buildgcc -P clang" will error out with the
"ERROR: Missing tool:" message if cmake is not installed.
Change-Id: I1e5c1bd67ade8f93ba0390df7f234deb47b9b18a
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/10556
Tested-by: build bot (Jenkins)
Reviewed-by: Francis Rowe <info@gluglug.org.uk>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
The precise phrase returned by 'type' differs between locales and shells.
It also doesn't matter because it returns an error code when it hasn't found a
match.
Let's simply assume there's no build_$OneOfOurPackages commands around that
could also match.
Change-Id: I44f021243149701e8da9dd74c368ca2ad4509419
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Tested-on: linux bash, linux dash, solaris sh, solaris ksh.
Reviewed-on: http://review.coreboot.org/10517
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
using grep is an extra process invocation, but it's not a bashism.
Also match precisely, so AGCC doesn't trigger on GCC (we don't have collisions
right now, but we won't have to deal with them in the future)
Change-Id: I242833c350b7f1e6a6793f288c1aae0b50d57a26
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Reviewed-on: http://review.coreboot.org/10518
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Build with bfd and gold linker, but use bfd linker per default
and make sure that lto is enabled in both binutils and gcc
Change-Id: I0584396b4580674cfdca24fbed0d8eeb1ee38806
Signed-off-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-on: http://review.coreboot.org/10496
Tested-by: build bot (Jenkins)
Reviewed-by: Edward O'Callaghan <edward.ocallaghan@koparo.com>
Instead of building IASL and GDB implicitly when building
GCC, this patch changes buildgcc to let you explicitly specify
what you want to build.
This will prevent IASL from building over and over again, when
all you need is GDB.
The new command line option is -P | --package <package> where
package is one of the following: GCC, GDB, CLANG, IASL
If no package is specified, buildgcc will default to GCC.
Change-Id: I8836bed16fc2bc39e0951199143581cc6d71cb4d
Signed-off-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-on: http://review.coreboot.org/10492
Tested-by: build bot (Jenkins)
Reviewed-by: Marc Jones <marc.jones@se-eng.com>
IASL was broken when compiling without GCC.
Change-Id: Ib859ce41c1dda10181781c025fc378504f5ebb91
Signed-off-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-on: http://review.coreboot.org/10495
Tested-by: build bot (Jenkins)
Reviewed-by: Edward O'Callaghan <edward.ocallaghan@koparo.com>
GDB stopped building ever since we updated from version 7.6
but nobody noticed ;)
Update from 7.9 to 7.9.1 and bring the required patches forward.
Change-Id: I2f357525a46d5540e9f57b80d830943bbd5dfcaf
Signed-off-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-on: http://review.coreboot.org/10494
Tested-by: build bot (Jenkins)
Reviewed-by: Edward O'Callaghan <edward.ocallaghan@koparo.com>
This groups all tasks happening in the main program,
orders them according to their dependencies and adds
comments on the various tasks.
Change-Id: Ib62bd213977cbc3307ef62e9a7e64515563968c1
Signed-off-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-on: http://review.coreboot.org/10490
Reviewed-by: Alexander Couzens <lynxis@fe80.eu>
Tested-by: build bot (Jenkins)
Patrick Georgi