Fix the following error from clang invoking gcc linker with wrong arg:
i386-elf-gcc: error: unrecognized command line option '--rtlib=libgcc';
did you mean '-static-libgcc'?
clang-4.0: error: linker (via gcc) command failed with exit code 1
Just remove --rtlib switch from CFLAGS relating to clang
Change-Id: Ife7ef6b6b47a04598fc67b40751bc59eed93b4af
Signed-off-by: Damien Zammit <damien@zamaudio.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/21354
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin Roth <martinroth@google.com>
For x64 and x86_32 configurations, apply the -march flag to both GCC and
Clang flags.
This solves the problem of Clang-compiled coreboot failing due to Clang
emitting SSE instructions for code that is executed while SSE is not
enabled.
This patch takes functionality targeted for GCC configurations and moves
it down a few lines, modifying CFLAGS instead of GCC_CFLAGS in order
that it applies to both GCC and Clang.
This is an alternate patch to CB:32887.
Signed-off-by: Alan Green <avg@google.com>
Change-Id: I6a6a6136b01a64d46f730ed19ebbeaadaf2183df
Reviewed-on: https://review.coreboot.org/c/coreboot/+/32923
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
The __pci_driver and __cpu_driver uses variable length arrays which are
constructed by the linker at build-time.
The linker always place the structs at 16-byte boundary, as per
"System V ABI". That's not a problem on x86, as the struct is exactly
16 Bytes in size. On other platforms, like x86_64 it breaks, because the
default data alignment isn't SysV compatible.
Set -malign-data=abi to make x86_64 gcc use the SysV psABI.
Fixes broken __pci_driver and __cpu_driver on x86_64.
Change-Id: I2491d47ed03dcfd8db110dfb181b2c5281449591
Signed-off-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/30116
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Similar to i686 on x86_32, compile for nocona on x86_64.
Nocona is the first Pentium 4 CPU that has long mode support.
Required for 64bit support.
Change-Id: Ied28f98f89610a748be8d66cf35814e9112a4407
Signed-off-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-on: https://review.coreboot.org/c/29877
Reviewed-by: Philipp Deppenwiese <zaolin.daisuki@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
POWER8 is a specific implementation of ppc64, which is by now outdated
(POWER9 has been on the market for a while). Rename arch/power8/ to
potentially cover a wider range of hardware.
TEST=Toolchains built before/after this commit can build coreboot for
emulation/qemu-power8 from before/after this commit.
Change-Id: I2d6f08b12a9ffc8a652ddcd6f24ad85ecb33ca52
Signed-off-by: Jonathan Neuschäfer <j.neuschaefer@gmx.net>
Reviewed-on: https://review.coreboot.org/c/29943
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Timothy Pearson <tpearson@raptorengineering.com>
Descriptions are taken from the files themselves or READMEs. Description
followed by a space with the language in marked up as code.
Change-Id: I5f91e85d1034736289aedf27de00df00db3ff19c
Signed-off-by: Tom Hiller <thrilleratplay@gmail.com>
Reviewed-on: https://review.coreboot.org/27563
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Philipp Deppenwiese <zaolin.daisuki@gmail.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
In the code we do the following in a number of places
to pre-initialize an array with a certain value before
overwriting some of the array with other values:
u8 mainboard_picr_data[FCH_INT_TABLE_SIZE] = {
[0 ... FCH_INT_TABLE_SIZE-1] = 0x1F,
}
clang does not like that behavior unless we specify
the option -Wno-initializer-overrides.
Remove the check for gcc in those places, too, because
1) it would silently change array contents between compilers
2) the check isn't sufficient to determine compilation on
clang vs gcc
Signed-off-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Change-Id: I93cc121b6fec099fcdbd5fd1114c2ff7cbc291dc
Reviewed-on: https://review.coreboot.org/20384
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
clang complains that the access might be unaligned. Yes, we know. Yes,
that's exactly what we want. You have _one_ job.
Change-Id: I5400f50d8b5b462270c700f7ff90d9d517278e71
Signed-off-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-on: https://review.coreboot.org/19659
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
That seems to be the more reliable way to build clang cross compilers
for now.
Change-Id: I14fe767d20f91b64e96c909291760bddcd108e5c
Signed-off-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-on: https://review.coreboot.org/19660
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Add a Kconfig value to enable building libpayload with the 586 compiler.
Update the cross compiler script to add the Kconfig value name that is
used when libpayload builds.
The Quark SOC does not support some of the instructions generated with
the 686 compiler (e.g. CMOV). Success occurs when
payloads/libpayload/build/config.h indicates that
CONFIG_LP_USE_MARCH_586=1.
TEST=Build and run on Galileo Gen2.
Change-Id: I04907e9a38ee139bae2e8b227821f54614707c25
Signed-off-by: Lee Leahy <Leroy.P.Leahy@intel.com>
Reviewed-on: https://review.coreboot.org/20322
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Martin Roth <martinroth@google.com>
clang, like gcc, needs a compiler runtime library. Unlike gcc, it can
use either its own runtime library (compiler-rt), or gcc's version
(libgcc). Also unlike gcc, the version of clang that is currently part
of our reference toolchain does not provide the necessary versions of
compiler-rt for all platforms we support. Hence, for now, use libgcc
even on clang builds. This patch allows switching between the two, but
switching to compiler-rt will break clang builds, unless someone fixes
our reference toolchain to provide libclang_rt.builtins-${ARCH}.a for
each of our supported platforms.
Change-Id: I5001a4b62ed34df19312f980b927ced8cbaf07db
Signed-off-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-on: https://review.coreboot.org/20303
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin Roth <martinroth@google.com>
The former only exists with a custom patch while the latter is supported
by clang and in the absense of libgcc even points to clang's own runtime
libraries.
Change-Id: I1e30d5518cf78e1d66925d6f2ccada60a43bb4f8
Signed-off-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-on: https://review.coreboot.org/19658
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Edward O'Callaghan <quasisec@google.com>
Reviewed-by: Martin Roth <martinroth@google.com>
Reviewed-by: Philippe Mathieu-Daudé <philippe.mathieu.daude@gmail.com>
Speed up the execution of this script from ~6 seconds to ~1 on my
system.
There are some changes to its output, but they're actually _more_
correct: so far, architectures without compiler support kept compiler
options for architectures that ran successfully earlier.
Change-Id: I0532ea2178fbedb114a75cfd5ba39301e534e742
Signed-off-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-on: https://review.coreboot.org/18262
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
This change adds armv7-r support for all stages.
armv7-r is an ARM processor based on the Cortex-R series.
Currently, there is support for armv7-a and armv7-m and
armv7-a files has been modfied to accommodate armv7-r by
adding ENV_ARMV7_A, ENV_ARMV7_R and ENV_ARMV7_M constants
to src/include/rules.h.
armv7-r exceptions support will added in a later time.
Change-Id: If94415d07fd6bd96c43d087374f609a2211f1885
Signed-off-by: Hakim Giydan <hgiydan@marvell.com>
Reviewed-on: https://review.coreboot.org/15335
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
In newer toolchain with binutils 2.26 and GCC 5.3.0, we build binutils
and GCC with machine type riscv32 and riscv64 instead of riscv. We can
see it in this riscv-gnu-toolchain commit:
https://github.com/riscv/riscv-gnu-toolchain/commit/dedbf07
Signed-off-by: Iru Cai <mytbk920423@gmail.com>
Signed-off-by: Martin Roth <martinroth@google.com>
Change-Id: Id552859ec256d80108e073d25cd51dd1fc3fbfac
Reviewed-on: https://review.coreboot.org/14257
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Just because an 'as' with a certain prefix is available does not guarantee
that a 'gcc' with the same prefix is available as well.
Without a check detect_compiler_runtime() would try to execute an
unavailable binary and print something like this:
.../xcompile: line 218: arm-linux-gnueabi-gcc: command not found
Change-Id: Icbadfeb2860152f7cf7696a9122521d0d881f3aa
Signed-off-by: Stefan Tauner <stefan.tauner@gmx.at>
Reviewed-on: https://review.coreboot.org/14563
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
This is taken from FILO and slightly enhanced.
Signed-off-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Change-Id: Ieadd9db3f1013ec1cd9f5a1dc44e17587617f1d1
Original-Change-Id: I961a7ddcd39657c9463806d7b82757eff0a4ac57
Original-Signed-off-by: Patrick Georgi <patrick.georgi@secunet.com>
Original-Reviewed-on: http://review.coreboot.org/190
Reviewed-on: https://review.coreboot.org/14386
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Do not disable warnings about unused but set variables to further
improve the code quality.
Change-Id: I25fa29ac42c9d09596d03f11fb01f31635a62a11
Signed-off-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-on: https://review.coreboot.org/3981
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Tested-by: build bot (Jenkins)
There was a report that xcompile wasn't finding the compilers correctly,
so to aid in future debugging, this adds a parameter to show what
xcompile is doing as it runs.
Run from the command line:
./util/xcompile/xcompile --debug
Change-Id: I779cb3de7b4e3f62a2ef2a6245c3538be518870c
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/13047
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Instead of instructing users to edit xcompile when they want to build
a quark platform, give the build a way to set -march=586 so that
the quark code will build correctly. The Quark processor does not
support the instructions introduced with the Pentium 6 architecture.
Change-Id: I0ed69aadc515f86f76800180e0e33bcd75feac5a
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/13552
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Tested-by: build bot (Jenkins)
Reviewed-by: Leroy P Leahy <leroy.p.leahy@intel.com>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-by: FEI WANG <wangfei.jimei@gmail.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>
Change-Id: Id0316042f665ec9c095887cf6a37a7949ed8e861
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Reviewed-on: https://review.coreboot.org/13421
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Tested-by: build bot (Jenkins)
Not just *-linux-gnu.
Change-Id: Ib817c6d207d3b69ce7595505f2b45f3be35b7d2f
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Reviewed-on: https://review.coreboot.org/13420
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Tested-by: build bot (Jenkins)
What's the exact difference between TARCH, TSUPP and TBFDARCHS? Fear no
more, it's documented.
Change-Id: I18717eb1e20b1c0a82a485d391de2794a77c59ae
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Reviewed-on: https://review.coreboot.org/13419
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Tested-by: build bot (Jenkins)
I think these four methods should cover most operating systems,
with many supporting several of the methods.
If we don't find anything, we're not any worse off than we were before.
The big issue would be if we get an incorrect value.
Change-Id: I4a612d39e93173e9d6e0de892f5bebf716912b1a
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/12937
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
I tried to handle the checking for the config flag internal to xcompile,
but the config flags don't appear to have been loaded into the
environment by make at that point.
This does update the if to check if the flag is even set before putting
anything into .xcompile though. If the LDFLAG isn't set, there's no
point in appending anything.
Also removes the LP version of the erratum config flag, which was a
copy/paste mistake from $(CONFIG_LP_COMPILER_GCC).
Change-Id: I3d8b0328c85310393a120741a498bc18867a6f54
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/12858
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Quoting variables prevents word splitting and glob expansion, and
prevents the script from breaking when input contains spaces, line
feeds, glob characters and such.
See shellcheck warning SC2086
Change-Id: Ib6ca46b64a621c4bea5c33ac312f2824b0386235
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/12845
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Using the local variables instead of positional parameters helps
readability.
- Add and use the local variables in testcc.
- Use the existing local variables in testld.
Change-Id: Ice13288b830a7aa043b360eaee8e36f060589a18
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/12844
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
While the backtick syntax isn't actually deprecated, the $() syntax
is preferred. Since both styles were being used in this script, settle
on the new standard for all cases.
Change-Id: I33770d666781b4fa34c909411e0d220c2540dbb4
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/12843
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Clean up the output file a bit by only including the erratum
for arm64 into the that architecture section instead of
every architecture.
Change-Id: Ib6276f12aee5deb92a03e1c4fa2ad57db46bdc8f
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/12842
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
In order to be able to check the compiler versions, we need to be
able to access the compiler variables. Move the original assignments
outside of the GCC check, and assign either the GCC or CLANG compiler
to the actual CC_ environment variable later. This ends up with the
same value set, while allowing the compiler versions to be checked.
Change-Id: Iffad02d526420ebbdfb15ed45eb51187caaa94fb
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/12841
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
We already have a CFLAGS variable - Use it for all of the flags.
Change-Id: I22b4c5cf24b8743e85ffab29ddcccdc6c732ea3b
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/12840
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
The XGCCPATH prefix is on all the other tools and compilers,
so add it to clang as well, so it can be found correctly.
Change-Id: Ibc250a81433f37bbb0555d32605aebe3a68aaf40
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/12839
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
- Add bar at the top of each architecture
- Make the architecture name and the TARCH_SEARCH to two lines
- Add a second line at the bottom of each architecture
- Add a comment about the two blank lines so they don't get
accidentally removed.
Change-Id: Ib4326bd94fe39b979244816ce54b752d083f6b16
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/12838
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Change-Id: I96a5048050f8016c3c569f20318b4d421a4470a7
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/12837
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Let toolchain.inc error out when the architecture or tool is missing.
Change-Id: I39a51e5a2c778d6bbc50354807e5e2b717fa9e52
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/12682
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Until there's a reason to, don't print a warning about the missing
power8 compiler.
Change-Id: I47c60e0a16892f0fa228e1439e0424926bca00a4
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/12634
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Alexander Couzens <lynxis@fe80.eu>
power8 is set up by ibm as a powerpc subset, so we follow
that rule here: we call it a powerpc but require -mcpu=power8
Change-Id: Ib5212be22db9584b0dc0eeed5c06ec1924347067
Signed-off-by: Ronald G. Minnich <rminnich@gmail.com>
Reviewed-on: https://review.coreboot.org/12624
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Bail out if .xcompile is incomplete or can't be regenerated.
Change-Id: I74adeded7a3e849b25bf65c5b02f67820f29c7e2
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Reviewed-on: http://review.coreboot.org/12477
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
On system with clang, "as" is available but "objdump" is not by default.
So if ${gccprefix} is empty, "as" can run successfully and the "objdump"
below might report error. Mask that output.
Change-Id: I9940f069f66e097973ed6138cf3c696087fa5531
Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Signed-off-by: Zheng Bao <fishbaozi@gmail.com>
Reviewed-on: http://review.coreboot.org/11681
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-by: Martin Roth <martinroth@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>
This patch adds support to enable a linker workaround to a hardware
erratum on some early Cortex-A53 revisions. Since the linker option was
added very recently, we use xcompile to test whether the toolchain
supports it first. It is also guarded by a Kconfig since only a few
ARM64 SoCs will need this and it incurs a performance penalty.
BRANCH=none
BUG=none
TEST=Turned it on or off for Smaug and confirmed that it (dis)appeared
in verbose make output accordingly.
Change-Id: I01c9642d3cf489134645f0db6f79f1c788ddb00d
Signed-off-by: Patrick Georgi <patrick@georgi-clan.de>
Original-Commit-Id: 57128785760c4dfa32d6e6d764756443a9323cb7
Original-Change-Id: Ia5dd124f484e38460d75fb864304e7e8b18d16b7
Original-Signed-off-by: Julius Werner <jwerner@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/294745
Original-Reviewed-by: Patrick Georgi <pgeorgi@chromium.org>
Reviewed-on: http://review.coreboot.org/11403
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-by: Julius Werner <jwerner@chromium.org>
Damien Zammit