With this commit we're starting a series of updates of ARC tools.
Significantly rewritten arc-2016.03 tools introduced way too many
problems highlighted by Buildroot autobuilder. Now in attempt to
resolve as many issues as possible by the time final release of
arc-2016.09 tools is cut we'll be executing arc-2016.09 series
with engineering snapshots like this one.
We decided to go this way instead of applying separate patches here
and there because ongoing development introduces quite a lot of
changes and separate patches are not practical in Buildroot.
Moreover this will give us very clean visibility of number of
issues we see (hopefully it will decrease over time).
One of the important changes introduced in this engineering build
is initial set of changes for proper support of PIE on ARC in terms
of both building on host and running on ARC target. I expect some
PIE-related build breakages to go away and new ones will be treated
as the high-priority issues to be fixed ASAP.
For now we only update Binutils and GCC while keeping GDB
as it is of arc-2016.03 release because there're some issues
we'd like to resolve before releasing it to wider audience.
So again note this is one of the first engineering builds of
arc-2016.09 series and it might have all kinds of breakages,
please don't use it for production builds.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: arc-buildroot@synopsys.com
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
binutils got disabled as a target package for aarch64 back in Buildroot
2012.11 (commit 660d16f4).
The default version 2.21.1 at the time didn't really support aarch64, as
shown by the autobuilder error reference in that commit.
binutils has now had support for aarch64 since 2.23, but was never
re-enabled for aarch64.
All of the binutils versions currently supported by Buildroot (oldest
being 2.24) support aarch64, so we can now enable it without any
restrictions.
This same patch also touches the dropwatch, lightning and oprofile
packages which all select binutils and had the 'depends on !aarch64' to
keep Kconfig consistent. They are now re-enabled for aarch64 as well.
They are changed in this same patch so that Kconfig is kept consistent
and doesn't end up having them disabled in a commit which only has the
binutils change applied.
lightning has other explicit arch dependencies so it is not actually yet
available in aarch64, but at least the dependency on !aarch64 because of
binutils is not needed anymore.
Everything has been build tested with Buildroot and external toolchains.
binutils and dropwatch work fine in a qemu target. oprofile doesn't seem
to be supported in the qemu aarch64 processor but builds fine and seems
to have some level of aarch64 support in the source, so there doesn't
seem to be a reason to keep it disabled in Kconfig.
Signed-off-by: Erico Nunes <nunes.erico@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Build failed when "makeinfo" was missing on the build host.
This was happening because "makeinfo" is required to build .info targets
and make exited with error. The issue hadn't appeared before as there was
prebuilt documentation in ARC binutils tarballs, so no attempts had been
made to build docs.
Missing "makeinfo" only stops us from building docs
("missing" script already throws a warning on that regard).
Let's continue to build other targets.
Now exit code of the script called "missing" is checked.
The value 127 means that "makeinfo" is not available on the build host.
So when such value occurs, 0 is returned to the top level makefile.
Documentation is not being built but further build of binutils continues.
Fixes http://autobuild.buildroot.net/results/55c/55cd09a559016f4f252f0e4c27313b9806135cf4//
Signed-off-by: Zakharov Vlad <vzakhar@synopsys.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Currently even if -pie flag is provided to LD dynamic relocations
won't be created. That in its turn will break U-Boot self-relocation
functionality.
This fix resolves mentioned problem.
Note as of today this is a sort of off-the-tree patch right from
our binutils guy but in coming day similar fix will be applied to
upstream binutils as well.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This change switches ARC tools to the most recent arc-2016.03
version.
ARC GNU tools of version arc-2016.03 bring some quite significant
changes like:
* Binutils v2.26+ (upstream commit id 202ac19 with additional ARC
* patches)
* GCC v4.8.5
* GDB 7.10
More about changes, improvements and fixes could be found here:
https://github.com/foss-for-synopsys-dwc-arc-processors/toolchain/releases/tag/arc-2016.03
Note in this change we're adding sha512 checksums for
both binutils and gcc tarballs fetched from GitHub.
Build and run-tested in nSIM for both ARC700 and ARC HS38.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Now that we have added support for the lastest binutils version 2.26,
it is time to remove the oldest version we currently support, 2.23, in
order to keep only 3 versions: 2.24, 2.25 and 2.26.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This change switches ARC tools to RC2 of the most recent arc-2016.03
version.
Essentially once final release is ready version will be bumped again.
ARC GNU tools of version arc-2016.03 bring some quite significant
changes like:
* Binutils v2.26+ (upstream commit id 202ac19 with additional ARC patches)
* GCC v4.8.5
* GDB 7.10
More about changes, improvements and fixes could be found here:
https://github.com/foss-for-synopsys-dwc-arc-processors/toolchain/releases/tag/arc-2016.03-rc2
Also in this change we realign custom Buildroot patches for binutils and
gcc for ARC tools. Looks like earlier most of arch-independent patches
for binutils and gcc were either unintentionally removed or not even
added in patch folders for ARC's binutils and gcc. Now arch-independent
patches for binutils-2.26 and gcc-4.8.5 were added in
package/{binutils|gcc}/arc-2016.03-rc2 folders.
Build and run-tested in nSIM for both ARC700 and ARC HS38.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
In commit 2f84595394 ("binutils: remove
version 2.22"), binutils 2.22 was removed, but forgot to remove the
corresponding hash from the hash file.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Plugins want to use dlopen even when configured with --disable-shared.
Add --disable-plugins when STATIC_LIBS is selected.
Fixes:
http://autobuild.buildroot.net/results/a59e5858d777dcee3d8f7fb6e82ee4735b520155/
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Recent commit 5eb37d6adc broke the build
by default to binutils 2.25 for the target, while we only have hashes
for 2.25.1. Use 2.25.1 instead.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
binutils 2.26 has been released, it is time to switch to 2.25 as the
default version.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Despite the documentation and the comment in xtensa_move_literals, in
the presence of --text-section-literals and --auto-litpools literals are
moved from the separate literal sections into .init and .fini, because
the check in the xtensa_move_literals is incorrect.
This fixes build errors seen with projects that have .init/.fini and use
text-section-literals.
Backported from: 4111950f363221c4641dc2f33bea61cc94f34906
Reported-by: Waldemar Brodkorb <mail@waldemar-brodkorb.de>
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
- Rebase patches from 2.25 on top of 2.26.
- Use git format for patches
- remove upstreamed xtensa patches
The first patch [1] is needed to be able to build glibc with binutils 2.26 on nios2.
The second from [2] fix an assertion issue with ld.
[1] https://sourceware.org/git/gitweb.cgi?p=binutils-gdb.git;a=commitdiff;h=a7be2893a6449e64fe6cfcdd8700b0a367a69f19
[2] https://sourceware.org/bugzilla/show_bug.cgi?id=19405
Signed-off-by: Romain Naour <romain.naour@gmail.com>
Cc: Gustavo Zacarias <gustavo@zacarias.com.ar>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Add the Cortex A17 variant. This core is considered a replacement
of the Cortex A12 and is supported by gcc 5 / binutils 2.25+
Suggested-by: Ross Green <greenfross@netscape.net>
Signed-off-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
When zlib was compiled before, binutils will use it as optional
dependency:
$ output/host/usr/bin/x86_64-linux-readelf -a output/target/usr/lib/libbfd-2.24.so | grep NEEDED
0x0000000000000001 (NEEDED) Shared library: [libz.so.1]
0x0000000000000001 (NEEDED) Shared library: [libc.so.1]
Apparantly the build system offers no support to control the
dependency, --without-system-zlib with zlib installed still provides
libbfd-2.24.so linked to libz.so.1.
Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Despite the documentation and the comment in xtensa_move_literals, in
the presence of --text-section-literals and --auto-litpools literals are
moved from the separate literal sections into .init and .fini, because
the check in the xtensa_move_literals is incorrect.
This fixes build errors seen with projects that have .init/.fini and use
text-section-literals.
Backported from: 4111950f363221c4641dc2f33bea61cc94f34906
Reported-by: Waldemar Brodkorb <mail@waldemar-brodkorb.de>
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This change introduces newer ARC toolchain in Buildroot.
That new arc-2015.12 release doesn't bring any significant changes
but mostly is focused on fixes and minor improvements here and there.
Most noticeable changes are:
* GCC: Source update to v4.8.5
* GDB: Updated to upstream 7.10 release.
You may find more info on fixes and improvements in that release at:
https://github.com/foss-for-synopsys-dwc-arc-processors/toolchain/releases/tag/arc-2015.12
Signed-off-by: Lada Trimasova <ltrimas@synopsys.com>
Cc: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Binutils runs the configure scripts of subdirs at make time, so we need a
custom build command to ensure TARGET_CONFIGURE_ARGS are taken into
consideration for those (E.G. for musl gettext handling).
Reported-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This change introduces newer ARC toolchain in Buildroot.
Note this is the first release candidate and we'll probably see another
RC before cutting the final release.
That new arc-2015.12 release doesn't bring any significant changes but
mostly is focused on fixes and minor improvements here and there.
Most noticeable changes are:
* GCC updated to v4.8.5
* GDB updated to 7.10
You may find more info on fixes and improvements in that release at:
https://github.com/foss-for-synopsys-dwc-arc-processors/toolchain/releases/tag/arc-2015.12-rc1
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Cc: arc-buildroot@synopsys.com
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
It turned out one of the previous fixes (required to build Linux
kernel)
------------------>8-------------------
ARC Binutils: a65b844aed
Buildroot: http://git.buildroot.net/buildroot/commit/?id=2d4e2e238a0ea9395152ae71d882d79b1f35094c
------------------>8-------------------
broke building of some other software packages.
In particular perf built for ARCv2 had corrupted .plt entries that lead
to immediate crash on perf execution.
That's an example of normal .plt entries:
------------------>8-------------------
< 1-st PLT entry >:
12c24: 30 27 8c 7f 0d 00 74 95 ld r12,[pcl,0x000d9574]
12c2c: 21 20 00 03 j.d [r12]
12c30: 0a 24 c0 1f mov r12,pcl
< 2-nd PLT entry >:
12c34: 30 27 8c 7f 0d 00 68 95 ld r12,[pcl,0x000d9568]
12c3c: 21 20 00 03 j.d [r12]
12c40: 0a 24 c0 1f mov r12,pcl
------------------>8-------------------
Note right after jump in its delay-slot r12 gets set with current value
of program counter. This is required for the first symbol resolution,
see implementation of _dl_linux_resolve here:
http://git.uclibc.org/uClibc/tree/ldso/ldso/arc/resolve.S#n46
And that's what we got in .plt after mentioned fixes:
------------------>8-------------------
< 1-st PLT entry>:
13384: 30 27 8c 7f 0f 00 84 75 ld r12,[pcl,0x000f7584]
1338c: 21 20 00 03 j.d [r12]
< 2-nd PLT entry>:
13390: 30 27 8c 7f 0f 00 78 75 ld r12,[pcl,0x000f7578]
13398: 21 20 00 03 j.d [r12]
------------------>8-------------------
Note r12 setup is missing.
That happened because linker thought the size of PLT entry is 12 bytes
(which is exactly the size of PLT entry for ARCv1, read ARC750/770)
while for ARCv2 PLT entry is 16-bytes long. And erroneously trailing 4
bytes were truncated.
Current commit fixes this misbehavior and PLT gets generated properly
again.
Now we have a fix for that issue, see
5df50c6108
This fix is in arc-2.23-dev branch and will be a part of the next
release of ARC tools, so then this patch must be removed from buildroot.
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
- Add support for mips32r6 and mips64r6 target architecture variants
- Disable unsupported gcc versions
- Disable unsupported binutils versions
- Disable unsupported external toolchains
- Disable unsuported C libraries
- Add a hook in order to make glibc compile for MIPS R6.
[Thomas: slightly tweak the glibc hack explanation, to make it
hopefully clearer.]
Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Auto-litpools is the automated version of text-section-literals: literal
pool candidate frags are planted every N frags and during relaxation
they are turned into actual literal pools where literals are moved to
become reachable for their first reference by L32R instruction.
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
host-texinfo dependency was only necessary when building the git version
of binutils, however, this upstream commit allows binutils to build
without makeinfo (a binary provided by host-texinfo)...
https://sourceware.org/git/gitweb.cgi?p=binutils-gdb.git;a=commit;h=bba33ab1e0f7d2ebd8f8435f92ed12e2a3c558a4
...so we can safely remove the host-texinfo dependency.
Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
With the fix for missing .tdata/.tbss sections we unintentionally
introduced a regression for ARCv2 ISA (read ARC HS38) kernel building.
That's what we got on attempt to build kernel:
----------------------------------->8--------------------------------------
LD drivers/video/fbdev/built-in.o
arc-linux-ld: ERROR: Attempting to link drivers/video/fbdev/omap2/built-in.o with a binary drivers/video/fbdev/built-in.o of different architecture
arc-linux-ld: failed to merge target specific data of file drivers/video/fbdev/omap2/built-in.o
scripts/Makefile.build:337: recipe for target 'drivers/video/fbdev/built-in.o' failed
make[3]: *** [drivers/video/fbdev/built-in.o] Error 1
scripts/Makefile.build:403: recipe for target 'drivers/video/fbdev' failed
make[2]: *** [drivers/video/fbdev] Error 2
scripts/Makefile.build:403: recipe for target 'drivers/video' failed
make[1]: *** [drivers/video] Error 2
Makefile:944: recipe for target 'drivers' failed
make: *** [drivers] Error 2
----------------------------------->8--------------------------------------
The reason was empty .tdata and .tbss sections in empty archives. And
later empty archives were linked in built-in.o with default architecture
(in our case ARCv1 ISA, read for ARC 700) and then expected failure
happened when objets for different architectures were attempted to link
together.
Now we have a fix for that issue, see
a65b844aed
This fix is in arc-2.23-dev branch and will be a part of the next
release of ARC tools, so then this patch must be removed from buildroot.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Two patches below are taken from development branch of ARC binutils,
so one the next release of ARC tools happens both patches must be removed.
These 2 patches:
0003-ld-arc-Provide-the-.tdata-symbol.patch
0004-Provide-.tbss-symbol.patch
fix missing .tbss and .tdata sections during linkage stage such as:
------------------------------>8---------------------------
arc-linux-g++ test.cpp
../arc-buildroot-linux-uclibc/lib/libstdc++.so: undefined reference to '.tbss'
collect2: error: ld returned 1 exit status
------------------------------>8---------------------------
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Rename the binutils configuration option to match that one used by gcc
where the patchlevel is explicitly left out.
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
I'm happy to update GNU tools for ARC cores to the most recent
arc-2015.06 release.
This release brings following major improvements:
* GCC: source update to v4.8.4
* GCC: C ABI compatibility between MetaWare and GNU toolchains
* uClibc: support for thread local storage and Native Pthread Library (NPTL)
* GDB: updated to version 7.9.1
Also a lot of fixes and improvements has been done, please refer to
https://github.com/foss-for-synopsys-dwc-arc-processors/toolchain/releases/tag/arc-2015.06
for more details.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Anton Kolesov <akolesov@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Cc: Romain Naour <romain.naour@openwide.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
We already use 2.24 as the default, 2.25 has been around for some
time, and 2.22 is clearly very old, so let's get rid of it.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
When binutils is not built for the host, binutils.mk decides of the
version to be used when building binutils for the target. We're still
using the old 2.22 binutils version, but it's time to bump up to 2.24,
which we already use as the default version for the host.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Building libgcc with TARGET_ABI flags results in assembler segfault on
xtensa, because code in sections .init and .fini emits literals w/o
.literal_position directive. This patch turns the segfault into assembly
error.
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Binutils 2.24 produces broken code when compiling the kernel for
ppc64le, so prevent this combination. See:
https://sourceware.org/ml/binutils/2013-12/msg00200.html
The problem manifests early in the boot process with "Kernel access of
bad area, sig: 11" in arch_match_cpu_phys_id().
The fix has been merged upstream as commit
57fa7b8c7e59e35bced580f9bcb9668af43fdbce, which is available since
Binutils 2.25.
Signed-off-by: Sam Bobroff <sam.bobroff@au1.ibm.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
In buildroot we maintain a couple off-the-tree patches.
In particular for binutils and gcc packages.
Because we have many versions of mentioned packages patches for each
particular version reside in a folder which name matches full version
name of the package.
For example we used to have patches for binutils distributed as part of
arc-2014.12 tools in folder ""package/binutils/arc-2014.12". Now with
bump of ARC tools version we need to rename folder with patches to
"arc-2015.06-rc1".
The same applies to gcc.
Should fix http://autobuild.buildroot.net/results/2b2/2b27a4a64c0b225ae479ecfccf7a97f5ea95598c/
As discussed before it's not possible to reproduce reported problem on recent
Fedora 21/22 distros (at least we know about them) but I may confirm that
patches were applied fine and everything was built well. Hopefully reported
build failure goes away now.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Romain Naour <romain.naour@openwide.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Following commit 36555b4c8d ("ARC:
switch to RC1 of upcoming arc-2015.06 tools"), the binutils package
only contains the hash information for the 2015.06-rc1 ARC version of
binutils (which is in fact no hash). However, when an external
toolchain is used, and binutils is built for the target, it's still
the old 2014.11 binutils version that gets used in binutils.mk,
causing a build failure because there is no hash available for this
version.
This commit fixes that by using 2015.06-rc1 as the default binutils
version on ARC, which is used when no host-binutils has been built.
Fixes:
http://autobuild.buildroot.org/results/8f7/8f772e6fccb4f918120a7bb814da2224432d1c09/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Even though this is only RC1 it's been heavily used internally so it should
not be any worse than existing arc-2014.12.
Moreover this relase (and so its RC1) finally delivers support of NPTL
for ARC in uClibc.
That's why it would be good to allow interested users to start trying it
(for example WebKit and apps that use WebKit could be successfully built
and run) also it will be helpful to run that new toolchain through
autobuilder in attempt to find any hidden regressions so we have a solid
toolchain for release.
If there's an interest in that patch more patches will follow with
subsequent RCs and essentially on appearence or relese Buildroot will be
updated with it.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Cc: arc-buildroot@synopsys.com
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
elf_xtensa_gc_sweep_hook doesn't correctly unreference symbols that were
made local, that results in link failure with the following message:
BFD (GNU Binutils) 2.24 internal error, aborting at elf32-xtensa.c line
3372 in elf_xtensa_finish_dynamic_sections
elf_xtensa_gc_sweep_hook determines symbol reference type (PLT or GOT) by
relocation type. Relocation types are not changed when symbol becomes
local, but its PLT references are added to GOT references and
plt.refcount is set to 0. Such symbol cannot be unreferences in the
elf_xtensa_gc_sweep_hook and its extra references make calculated GOT
relocations section size not match number of GOT relocations.
Fix it by treating PLT reference as GOT reference when plt.refcount is
not positive.
Fixes:
http://autobuild.buildroot.net/results/3e2e24f697e26c93d4d95782b1cb7799fa620a7a/http://autobuild.buildroot.org/results/97d4c96d6f6cdc1ed4007456f4ab70be9dfa41b5/
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
--disable-install-libiberty configure option is broken
in gcc 4.8.x, so libiberty.a is always installed in HOST_DIR.
This library broke the host-gdb build due to a fpic/fPIC issue.
Since binutils use a copy of gcc's libiberty sources, it's
affected by the same bug.
binutils-arc-2014.11 which is based on 2.23 branch doesn't have
the PR56780 fixies, so backport them.
Fixes:
http://autobuild.buildroot.net/results/ca4/ca45297aa0ffbc9062ed92dc7ac070b0b33001de/
Signed-off-by: Romain Naour <romain.naour@openwide.fr>
Cc: Alexey Brodkin <abrodkin@synopsys.com>
Acked-by: Alexey Brodkin <abrodkin@synopsys.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
xtensa trampolines relaxation optimization caused the following build
errors:
Error: operand 1 of 'j' has out of range value '131643'
Error: operand 1 of 'j' has out of range value '4294836162'
Error: operand 1 of 'j' has out of range value '4294836072'
Extra condition 'abs (addr - trampaddr) < J_RANGE / 2' for trampoline
selection results in regressions: when relaxable jump is little longer
than J_RANGE so that single trampoline makes two new jumps, one longer
than J_RANGE / 2 and one shorter, correct trampoline cannot be found.
Drop that condition.
Upstream status: patch submitted.
Fix squashed into the optimization patch.
Fixes:
http://autobuild.buildroot.net/results/5ba7d18262ce6a2dfd69db07d064a971267f1128/
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This optimization alone reduces gnuradio build time for xtensa from 2
hours to 40 minutes.
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This series optimizes most time-consuming algorithms and data structures
in the xtensa link-time relaxation code, leaving relaxation logic intact.
Speedup linking typical linux kernel is ~8 times (1 minute instead of 8),
pathological cases (linking objects partially linked without relaxation)
are handled ~60 times faster (1 minute instead of an hour).
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Now that we don't support the internal blackfin toolchain any more
remove unnecessary conditionals.
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
When an external toolchain is used on ARC, BR2_BINUTILS_VERSION is
undefined, so we use binutils 2.22, which doesn't support ARC and
anyway isn't available from the ARC download location.
So, let's defined a default ARC binutils version in binutils.mk.
Fixes:
http://autobuild.buildroot.org/results/52a/52abadacd7aab2d5c11d094937f28198bf220662/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Add a new option BR2_GCC_ENABLE_LTO that builds gcc and binutils with
LTO support.
Individual packages still have to enable LTO explicitly by passing '-flto' to
GCC, which passes it on to the linker. This option does not add that flag
globally. Some packages detect if the compiler supports LTO and enable the flag
if it does.
To support LTO, ar and ranlib must be called with an argument which triggers the
usage of the LTO plugin. Since GCC doesn't call these tools itself, it instead
provides wrappers for ar and ranlib that pass the LTO arguments. This way
existing Makefiles don't need to be changed for LTO support. However, these
wrappers are called <tuple>-gcc-ar which matches the pattern to link to the
buildroot wrapper in the external toolchain logic. So the external toolchain
logic is updated to provide the correct symlink.
[Thomas:
- Add a separate BR2_BINUTILS_ENABLE_LTO option to enable LTO
support in binutils. This is a blind option, selected by
BR2_GCC_ENABLE_LTO. It just avoids having binutils.mk poke
directly into gcc Config.in options.
- Remove the check on the AVR32 special gcc version, which we don't
support anymore.
- Adapt the help text of the LTO Config.in option to no longer
mention "Since version 4.5", since we only support gcc >= 4.5 in
Buildroot anyway.
- Fix typo in toolchain-external.mk comment.]
Signed-off-by: Peter Kümmel <syntheticpp@gmx.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Versions previous to binutils v2.25 do not support the Nios-II architecture,
so disable them.
Signed-off-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
