For nativesdk, symlinks created by do_install contain ${SDKPATHNATIVE}.
This is a problem with SDKs that include nativesdk-binutils and nativesdk-gcc.
When extracting such an SDK dangling symlinks are created, because
relocate_sdk.py does not adjust symlinks. As a result gcc ends up calling the
host binutils.
Use the os.path.relpath function to obtain shorter relative paths, which do not
contain ${SDKPATHNATIVE}.
(From OE-Core rev: 3d6372aa2d9f752eac0e4943d1bde04f0f1924a0)
Signed-off-by: Achille Fouilleul <achille.fouilleul@gadz.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
A recent binutils patch added the LDGOLD variable but its already used
for controlling EXTRA_OECONF options for gold. Separate the two variables
to different names to avoid build warnings and confusion.
(From OE-Core rev: 713c83fd00ab16250f05b0c3933f0c8178b8a33e)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This is needed for riscv64 to build target binutils fixes alternatives
creation
(From OE-Core rev: 81a8e7addadcffc322735f3a30369d382334590b)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When the assembler reports that the input and output are the same, report the
file names involved, in order to help debugging. Also do not equate two files
are the same if the have the same inode value but reside on different file
systems.
(From OE-Core rev: 83cb0938b90bab9ba727f883b8955b0b40d49a01)
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Add x86_64-pep emulation support to the set enabled for x86_64 targets
to enable the linker to produce Portable Executables for EFI binaries.
Enables building the x86-64 EFI variant of the Xen hypervisor for
the OpenXT Project.
(From OE-Core rev: 3d493838f0284948a09693ad65656cb7afbb21cc)
Signed-off-by: Christopher Clark <christopher.clark6@baesystems.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Additionally cherry-pick
5ffeba4a09 Revert "PowerPC PLT speculative execution barriers"
b01452b1d4 [PR22764][LD][AARCH64]Allow R_AARCH64_ABS16 and R_AARCH64_ABS32 against absolution symbol or undefine symbol in shared object.
a985e9b9de Import patch from mainline to remove PROVODE qualifiers around definitions of __CTOR_LIST__ and __DTOR_LIST__ in PE linker scripts.
eec4607fc5 Add support for DWARF-4 line number tables.
(From OE-Core rev: c708506eb9dbb4b817f563fbaacb80eee0b5b301)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This makes it easy to override them in bbappends
(From OE-Core rev: 99458d1fa170970e8a1aa04eeb0c86d3754fd988)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Whenever SDKMACHINE is set to mingw32, sdkmingw32 override is defined
everywhere. This meant that value of LDGOLD was different also for
binutils and binutils-cross depending whether SDKMACHINE was set or not.
(From OE-Core rev: 0398f84482dbe5ee99b20e2ca6b3e9984f7ccec5)
Signed-off-by: Samuli Piippo <samuli.piippo@qt.io>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Update to the latest commit on the 2.28 branch to pick up:
a3e00a8 [GOLD] Avoid duplicate PLT stub symbols on ppc32
along with several backported bug fixes for arm*, etc.
This resolves the systemd link error for qemuppc
with the error message:
.../powerpc-oe-linux-ld.gold:
error: invalid STB_LOCAL symbol in external symbols
error: linker defined: multiple definition of '00000001.plt_call.memcpy+8000'
command line: previous definition here
described in:
[YOCTO #11696]
(From OE-Core rev: 0d1273cdf8652bf1844f1f1fe4ba0eacd27ada65)
Signed-off-by: Randy MacLeod <Randy.MacLeod@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Some of these are clearly dead, e.g. one binutils patch reverts the effects
of the earlier one.
This also removes the uclibc site files. We now have mechanisms to allow these
to be extended from another layer should someone ever wish to do that.
(From OE-Core rev: e01e7c543a559c8926d72159b5cd55db0c661434)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Drop patches to ChangeLog, they are in patch
header anyway
(From OE-Core rev: 0b0f545dbf16b0970c5a79975d451dc9d887c2a7)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Some tools (my issue was with 'perf') only need the libbfd component of
binutils, so we can save space in images by splitting it out.
Previously, instead of `perf` pulling in `libbfd-VERSION.so`, it pulled
in all the libraries and binaries from binutils.
(From OE-Core rev: d57109e44f1a32d0e39d250ad7e71d7d80757d35)
Signed-off-by: Cody P Schafer <dev@codyps.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
"yocto-compat-layer.py --machines" showed that shared packages like
gcc-cross-powerpc64 have a sstate signature that depends on
TUNEFLAGS. As a result, there are unnecessary rebuilds and potential
conflicts in a multiconfig.
That's due to the way how TARGET_ARCH is set. Richard Purdie suggested
setting TARGET_ARCH[vardepvalue] as fix, which works. It would be
shorter to do that in cross.bbclass instead of repeating the relevant
line in different recipes, but Richard was concerned about potential
side-effects in other usages of cross.bbclass.
TARGET_GOARM as used in go.inc is still causing signature differences
for go-cross-powerpc64 and machines b4420qds-64b and p5020ds-64b. This
needs further investigation.
(From OE-Core rev: 39bfa0dd3237cbca47e7fca1075d521f9d073f25)
Signed-off-by: Patrick Ohly <patrick.ohly@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Backport upstream commit to address vulnerabilities:
CVE: CVE-2017-6965
[BZ 21137] -- https://sourceware.org/bugzilla/show_bug.cgi?id=21137
Fix readelf writing to illegal addresses whilst processing corrupt input
files containing symbol-difference relocations.
PR binutils/21137
* readelf.c (target_specific_reloc_handling): Add end parameter.
Check for buffer overflow before writing relocated values.
(apply_relocations): Pass end to target_specific_reloc_handling.
CVE: CVE-2017-6966
[BZ 21139] -- https://sourceware.org/bugzilla/show_bug.cgi?id=21139
Fix read-after-free error in readelf when processing multiple, relocated
sections in an MSP430 binary.
PR binutils/21139
* readelf.c (target_specific_reloc_handling): Add num_syms
parameter. Check for symbol table overflow before accessing
symbol value. If reloc pointer is NULL, discard all saved state.
(apply_relocations): Pass num_syms to target_specific_reloc_handling.
Call target_specific_reloc_handling with a NULL reloc pointer
after processing all of the relocs.
(From OE-Core rev: 8c52a530ba2beb438aa47956bcec3777a1eafe5f)
Signed-off-by: Yuanjie Huang <yuanjie.huang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
oe-core 759eed (binutils: Enable threading when gold is enabled and is not
default linker) causes linking in mingw SDKs to fail:
.../work/i686-nativesdk-mingw32-pokysdk-mingw32/binutils-cross-canadian-x86-64/2.28-r0
/recipe-sysroot-native/usr/bin/i686-pokysdk-mingw32/../../libexec/i686-pokysdk-mingw32/gcc/i686-pokysdk-mingw32/6.3.0/ld:
cannot find -lpthread
Work around this by disabling gold entirely in mingw SDKs.
(From OE-Core rev: b5a595a4be09756b88e91f3353e3b221b165ab44)
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Currently we enable threaded linking feature of gold linker only
when its used as default ld. There is no need to restrict it when
its not default linker either. As long as gold is enabled, which
is the case here, we should be able to do threaded linking.
(From OE-Core rev: 759eed2b02e0a7b5c8b19d4b087d9151c009eed4)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We don't put target libs into a native/cross ${exec_prefix} but having
this in the default search path means all linker scripts have to be relocated.
This is a considerable chunk of files to create multiple copies of for no good
reason.
Instead, patch out the paths we don't need.
(From OE-Core rev: 20816eb5398512652c971a37589a2ca28ffd3d68)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There was bug with alignment frags for aarch64 in binutils. This is fixed in
master of binutils. This patch backports the fix to binutils 2.27 version.
(From OE-Core rev: f6f87019073d4f3caa7766aca89faa6781690fba)
Signed-off-by: Manjukumar Matha <manjukumar.harthikote-matha@xilinx.com>
Signed-off-by: Nathan Rossi <nathan@nathanrossi.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
getVar() now defaults to expanding by default, thus remove the True
option from getVar() calls with a regex search and replace.
Search made with the following regex: getVar ?\(( ?[^,()]*), True\)
(From OE-Core rev: 7c552996597faaee2fbee185b250c0ee30ea3b5f)
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The following patch fix build for c293pcie PPC machine :
https://sourceware.org/git/gitweb.cgi?p=binutils-gdb.git;a=commitdiff;h=8941017b
This patch is a backport from master branch.
Fix error:
/
|ERROR: binutils-2.27-r0 do_package: runstrip: ''powerpc-fslc-linux-gnuspe-strip'
|--remove-section=.comment --remove-section=.note --strip-unneeded
|'../tmp/work/ppce500v2-fslc-linux-gnuspe/binutils/2.27-r0/package/usr/lib/
|libbfd-2.27.0.20160806.so'' strip command failed with 1
|(b'powerpc-fslc-linux-gnuspe-strip: ../tmp/work/ppce500v2-fslc-linux-gnuspe/
|binutils/2.27-r0/package/usr/lib/stJMAEnm: Not enough room for program headers,
|try linking with -N\npowerpc-fslc-linux-gnuspe-strip:../tmp/work/
|ppce500v2-fslc-linux-gnuspe/binutils/2.27-r0/package/usr/lib/stJMAEnm
|[.note.gnu.build-id]: Bad value\n')
\
(From OE-Core rev: 3526c23ae98d2e08af20b77175c12907355ec42a)
Signed-off-by: Fabio Berton <fabio.berton@ossystems.com.br>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Fixes segfaults when doing partial linking and generaring binary objects
/tmp/binu/ld/ld-new -r -b binary -o etc_certs_ui_ca_pem.o etc_certs_ui_ca_pem
0 elf32_arm_count_additional_relocs (sec=0x79bf40) at /mnt/a/work/oe/binutils-gdb/bfd/elf32-arm.c:18210
1 0x000000000047635a in bfd_elf_final_link (abfd=abfd@entry=0x783250, info=info@entry=0x748400 <link_info>) at /mnt/a/work/oe/binutils-gdb/bfd/elflink.c:11224
2 0x000000000044df7b in elf32_arm_final_link (abfd=0x783250, info=0x748400 <link_info>) at /mnt/a/work/oe/binutils-gdb/bfd/elf32-arm.c:12131
3 0x0000000000418917 in ldwrite () at /mnt/a/work/oe/binutils-gdb/ld/ldwrite.c:577
4 0x000000000040365f in main (argc=<optimized out>, argv=<optimized out>) at /mnt/a/work/oe/binutils-gdb/ld/ldmain.c:433
gold works ok. The patch is already applied in master binutils
(From OE-Core rev: 00d1913520f1572fa7def865e57852c7f25b0ec4)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Warn system directories requires a configuration option to be enabled.
This patch enables the warning for cross-canadian ld, to align with the
cross version.
(From OE-Core rev: 581a746f081a265e08b41712e7bc953f46c6e92a)
Signed-off-by: Yuanjie Huang <yuanjie.huang@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We don't autoreconf/libtoolize binutils as it has very strict requirements, so
extend our patching of the stock libtool to include two fixes to RPATH
behaviour, as part of the solution to ensure that native binaries don't have
RPATHs pointing at the host system's /usr/lib.
This generally doesn't cause a problem but it can cause some binaries (such as
ar) to abort on startup:
./x86_64-pokysdk-linux-ar: relocation error: /usr/lib/libc.so.6: symbol
_dl_starting_up, version GLIBC_PRIVATE not defined in file ld-linux.so.2 with
link time reference
The situation here is that ar is built and as it links to the host libc/loader
has an RPATH for /usr/lib. If tmp is wiped and then binutils is installed from
sstate relocation occurs and the loader changed to the sysroot, but there
remains a RPATH for /usr/lib. This means that the sysroot loader is used with
the host libc, which can be incompatible. By telling libtool that the host
library paths are in the default search path, and ensuring that all default
search paths are not added as RPATHs by libtool, the result is a binary that
links to what it should be linking to and nothing else.
[ YOCTO #9287 ]
(From OE-Core rev: 6b201081b622cc083cc2b1a8ad99d6f7d2bea480)
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There was a clear typo in a function name, correct it.
(From OE-Core rev: dcf44e184a807d76463a3bf1b2315e80b9469de3)
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When changing SDKMACHINE, we may encounter an error forcing us to wipe the TMP folder.
Since only SDK_ARCH is captured in the PN of the crosssdk recipes, changes to SDK_OS
result in conflicts. Eventually we hit the error:
ERROR: ...: The recipe <...> is trying to install files into a shared area when those files already exist.
The build has stopped as continuing in this scenario WILL break things
This patchset addresses the problem by SDK_SYS as the recipe name suffix instead
of SDK_ARCH.
[YOCTO #9281]
(From OE-Core rev: d2eccccb70e809d482c493922f23aef4409cfd82)
Signed-off-by: Juro Bystricky <juro.bystricky@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Problem described here
https://lists.gnu.org/archive/html/bug-gettext/2015-11/msg00012.html
gettext does not detect the gettext support in libc
correctly if the libc is not glibc. Musl does support
the gettext version 1 and 2 of APIs
http://www.openwall.com/lists/musl/2015/04/16/3
tests in gettext.m4 however fail since it pokes at glibc
internal symbols to determine the gettext APIs
musl's implementaitons are done differenty so the
tests fail and hence it does not enable the libc
implementation. Since we install the header from
libc it confuses the compilation and results in errors
like
libbfd.so: undefined reference to `libintl_dgettext'
see
http://savannah.gnu.org/bugs/?46436
binutils need these variables in make env since
binutils build system runs configure in the sub directories
during make step, so we need to pass these flags
in compile step in addition to configure step
(From OE-Core rev: 21bba0548463f277684cc52d23194ad6d7c17956)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
As exposed by WebKit on aarch64 hosts, which causes binutils to throw an
internal error.
[ YOCTO #9509 ]
(From OE-Core rev: b31294e4f34dfb530c40526ab56c07aedb76e31b)
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This fixes the following compilation error when building a mipsel
yocto kernel for qemu:
| CC arch/mips/mm/sc-ip22.o
| {standard input}: Assembler messages:
| {standard input}:128: Error: number (0x9000000080000000) larger than 32 bits
| {standard input}:151: Error: number (0x9000000080000000) larger than 32 bits
| {standard input}:186: Error: number (0x9000000080000000) larger than 32 bits
We leave out the testsuite bits and the changelog in this
backport.
(From OE-Core rev: 8b378a17bf6d6c43f097b9df491e5c6ec59bf316)
Signed-off-by: André Draszik <adraszik@tycoint.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The wait mnemonic for ppc targets is incorrectly assembled into 0x7c00003c due
to duplicated address definition with waitasec instruction. The issue causes
kernel boot calltrace for ppc targets when wait instruction is executed.
(From OE-Core rev: 9764de92d5673d0f629555723321c933db015fe0)
Signed-off-by: Zhenhua Luo <zhenhua.luo@nxp.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
It's disabled on cross builds, and it's needed for gcc 6
(From OE-Core rev: ce1b37e29dc89b67dc698e856007b59faa16c4df)
Signed-off-by: Dan McGregor <dan.mcgregor@usask.ca>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
elfedit and readelf contains /usr/lib in elf header
this patch deletes them from binaries, ideally it
should be fixed in libtool and Makery of binutils
mips target binutils dont build gold so remove
them from ALTERNATIVES list
depend on own version of chrpath native, so builds on
build OS like Centos can work, the verison of chrpath
on centos is old enough to not support dealing with
multi-arch ELF files.
(From OE-Core rev: 9043202f4e705932a3847f2f7635fca06fe5a916)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
