Patches explain the issue in detail but this is exposed
with gcc 4.9 in binutils 2.23.2
(From OE-Core rev: fc5c467b680fc5aef4b0f689e6988e17a9322ae0)
(From OE-Core rev: 4dfb8847ebf8aab90ad8888933468e2899c96998)
(From OE-Core rev: af347d3298e15552d502d5b2ce497bbda9705bc7)
(From OE-Core rev: 07a7228392ec5157616888cee1eb119f4adb39a7)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Commit 616354f13732d13c17434d5b60b166f691c25761 is insufficient because
gnu-config-native's gnu-configize script uses perl modules from autoconf
and hence doesn't work unless autoconf-native is staged (which it may
not be if building from sstate).
Ideally g-c-n would itself declare a dependency on autoconf-native but this
is difficult to arrange without creating a dependency loop. autoconf-native
already depends on gnu-config-native (because autoreconf invokes gnu-configize)
and has a build dependency on m4-native, which in turn build-depends on g-c-n
because it configizes itself by steam in do_configure and needs config.{guess,sub}
to be available. Adding some sort of gnu-config-initial-native recipe would
fix the latter problem, but this would be ugly because it would need special-casing
in (at least) autotools.bbclass, and in any case this still wouldn't solve
the problem of autoconf itself depending on g-c-n.
So, the easiest solution to the problem at hand is to arrange for those
few recipes that depend on g-c-n but not autoconf-native to gain that
latter dependency as well.
(From OE-Core master rev: 507199e57acfcc99639dc2c53abe194d77d60866)
(From OE-Core rev: bbf8f596ca51aa33bdb5b0d5664827d62408863c)
Signed-off-by: Phil Blundell <pb@pbcl.net>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
do_configure() in binutils.inc includes an explicit call to
gnu-configize so we need to make sure that gnu-config-native is
present. Previously this was being dragged in with the rest of the
autotools stuff, but commit 54a3e2ee37003fc56af0339f857b0b6442790c26
disabled that for binutils-cross on the grounds that "we don't
autoreconf" the toolchain components. Fix this by adding
gnu-config-native itself explicitly to DEPENDS.
(From OE-Core master rev: 616354f13732d13c17434d5b60b166f691c25761)
(From OE-Core rev: c7b50b95ab27d3ea3b37e3ee16050af962746f5c)
Signed-off-by: Phil Blundell <philb@gnu.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
PowerPC toolchains can use the OS "linux" or "linux-gnuspe". This
patch links them together so the one cross-canadian toolchain can support
both.
GCC_FOR_TARGET is set for the GCC recipe as otherwise configure
can pick up an incorrect value.
[YOCTO #5354]
(From OE-Core master rev: a1d6331238982b0c5d39b0a18794f6654b00d46a)
(From OE-Core rev: b114b045687776ebc5c805e1f19758e7e37eebf2)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The cross-canadian compilers are now build once per architecture but were being
installed into tune specific locations which is incorrect. This adjusts things
so they are make TARGET_ARCH specific. We gain the tune specific parts from the
target sysroot which remains tune specific, the compiler and tools are independent
ot that.
binutils/gcc require sysroot options but since we reset at runtime, these shouldn't
have dependencies in the sstate checksums. They are therefore also excluded.
With these patches, switching machines does not result in a rebuild of *-cross-canadian
and the compiler is correctly located and referenced in the target images.
(From OE-Core rev: f58acab6414fe96d9e07ebbe86b348d2ac2bed5f)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Back port two patches to fix the error of complex expressions before @l/@h.
Error sample:
test.s: Assembler messages:
test.s:1: Error: operand out of range (0xffffffffffff8000 is not between 0x0000000000000000 and 0x000000000000ffff)
Fix Bug 4524.
(From OE-Core rev: 9ba812ab1f613d28f9eb3192d2ff1a34dfce33e4)
Signed-off-by: Zhenhua Luo <zhenhua.luo@freescale.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Update required to pass "movi" related build errors when gcc-4.8 is
used.
libgcrypt, slang, mysql5 were failing like this:
| {standard input}: Assembler messages:
| {standard input}:316: Error: immediate value out of range 0 to 255 at operand 2 -- `movi v3.8b,-106'
| {standard input}:348: Error: immediate value out of range 0 to 255 at operand 2 -- `movi v3.8b,-8'
| {standard input}:352: Error: immediate value out of range 0 to 255 at operand 2 -- `movi v3.8b,-27'
(From OE-Core rev: 2489151dbfc8bc002d89ab199d457ab3794c54a8)
Signed-off-by: Marcin Juszkiewicz <marcin.juszkiewicz@linaro.org>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There is a race condition between when the CONFIG_SITE list is generated and
then used here via the cat to create the config.cache and in *all* configure
scripts when reading from CONFIG_SITE.
The race in this case is that the sstate setscene task runs on a package that
contains a site config file (ncurses in this case) and then bitbake decides
that it needs to rebuild and cleans out the site config file, so it existed
for siteinfo_get_files() to find in SITECONFIG_SYSROOTCACHE and then was removed
for the rebuild. When bintuils tried to run the do_configure() task which creates
the binutil's version of config.cache it reads from CONFIG_SITE which now contains
the non-existant site config file. (confused yet ;-)?
Currently the configure script does a test -r to ensure the file is readable
before using it, therefore having the cat succeed regardless of the file being
available is consistent behaviour.
(From OE-Core rev: ffd8f05e5548500199c9b04a174067811ad2c5e7)
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This is a backport of a whitespace skipping issue: PR14987 and PR14887
(From OE-Core rev: f4312eaeb6eb2b7196be2b968e2e468e705fb2cf)
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This can cause issues building ppc64/eabi kernel. For details
see the patch header
(From OE-Core rev: 3c6622e22b0e1b2a8ceea6465ea84c6fb8299518)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
binutils build fails on Fedora18+:
1. binutils-2.23.1/bfd/elf32-xtensa.c:6078:36: error: argument to 'sizeof' in 'memset' call is the same expression as the destination; did you mean to dereference it? [-Werror=sizeof-pointer-memaccess]
memset (sec_cache, 0, sizeof (sec_cache));
^
2. binutils-2.23.1/bfd/elf32-xtensa.c:6120:32: error: argument to 'sizeof' in 'memset' call is the same expression as the destination; did you mean to dereference it? [-Werror=sizeof-pointer-memaccess]
memset (sec_cache, 0, sizeof (sec_cache));
^
3. binutils-2.23.1/opcodes/arc-dis.c:430:13: error: argument to 'sizeof' in '__builtin_strncat' call is the same expression as the destination; did you mean to provide an explicit length? [-Werror=sizeof-pointer-memaccess]
sizeof (state->commentBuffer));
^
4. binutils-2.23.1/opcodes/rl78-dis.c:230:13: error: array subscript is above array bounds [-Werror=array-bounds]
if (oper->use_es && indirect_type (oper->type))
^
(From OE-Core rev: 5445e12e5a32cc5c51ce8a29f2800692ed831115)
Signed-off-by: Zhenhua Luo <zhenhua.luo@freescale.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Drop virtclass override in favor of class override
(From OE-Core rev: b2e2cf532ebbeb7463b45193fa684e785e299c99)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When building the SDK, the final .interp section size should be set to a
bigger value (0x1000) in order to be able to change the dynamic loader's
path later. However, we shouldn't do that for partial linked objects
(when -r or -rU is used). That's because those objects will then have an
.interp section of 0x1000 even if it contains no data and when the final
linking is done we will end up with a "cannot move location counter
backwards" error. That's because the linker will try to squeeze all the data in
the .interp sections found in various partial linked objects into one 0x1000
bytes final .interp section.
[YOCTO #3264]
(From OE-Core rev: b25d0c5fe286e44ded46aefdcbe35ed259087759)
Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There does not exist an '--enable-target=all' option
(From OE-Core rev: 60fe4e80ca5845a0d03f918b80d6e980c13378b9)
Signed-off-by: Enrico Scholz <enrico.scholz@sigma-chemnitz.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The overrides virtclass-native and virtclass-nativesdk are deprecated,
which should be replaced by class-native and class-nativesdk.
[YOCTO #3297]
(From OE-Core rev: bb67ddeb2eed3e25c626a279ef53a7e8c7bfe6f2)
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
binutils will build differently if this feature is enabled, so
make the do_configure step depend on it
(From OE-Core rev: 0788cf349fe37ef4a36c626dbc396c97d1ab14d7)
Signed-off-by: Matthew McClintock <msm@freescale.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This patch has been causing some regressions on gold.
e.g. systemd based images segfault and uclibc based images
dont boot. There has been few other reports on the mailing
list. Considering this lets withdraw this patch.
(From OE-Core rev: ecbe671de1553956f83798e1c6fa3ec2fc6a7b4e)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
These are fixes mainly cherrypicks for mips/ppc/x86
mainly fixing PRs in ld and gold
(From OE-Core rev: f098cfc24bae8e0685bcae53ea4fdc3326ddc6c4)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
As discussed on the mailing lists, using a suffix to package names is
hard and has lead to many recipes having to do PKGSUFFIX games. Its
looking extremely hard to scale nativesdk much further without hacking
many recipes.
By comparison, using a prefix like multilib does works much better and
doesn't involve "hacking" as many recipes. This change converts nativesdk
to use a prefix using the existing multilib infrastructure.
(From OE-Core rev: 81813c0e322dc04ce4b069117188d8a54dfddb8c)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The -rpath argument would search the host filesystem for libraries,
even when a sysroot was defined. For cross toolchains with targets
compatible with the host architecture this can find incorrect
libraries. Leave -rpath-link unmodified, as build systems in the wild
are already using this to point to host directories.
[YOCTO #2965]
(From OE-Core rev: bccea580f1abb762d231f785a4e60c9cd368dcdf)
Signed-off-by: Andy Ross <andy.ross@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This options will let gold spawn multiple
threads for linking and speeding up linking
on multicore build hosts.
(From OE-Core rev: 210957a58260ec50416649aba799d0e96aeeff24)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This patch is needed so that all SDK binaries have the PT_INTERP section
size set to 4096 (max path size in Linux) in order to be able to parse
the binaries later and change the interpreter to the path we want.
(From OE-Core rev: 9bf657bbdfe7c1a9dff1cc121eab496e3e407197)
Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
e.g. libfl from flex is needed for target
binutils to link otherwise it can fail to build
(From OE-Core rev: 627725700ee9e2ca39a19f19a77cd772cd8ecf9f)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This updates the E5500 and E6500 support patches
as they have been applied upstream binutils.
(From OE-Core rev: 202420871785cbdbbf57adbe26eb0f649e57512b)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Plugins work well in 2.22+ so lets enable them
helps with lto
(From OE-Core rev: 81ecc9d468c36b36c4b2ccc4ab5f366611a8fa46)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Also rearrange the recipes to have common bits
in inc files and not include the target bb file
everywhere. This lets us add specific options
to specific recipes particularly target recipe in
this case
(From OE-Core rev: f4ed063e32f064e996a4c29760fa4ac49f1ed73c)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This is reported by prelinker which does not
go well with textrels.
(From OE-Core rev: 7b1caeca8e5c0dd066bb940fe6d2dff4d22bf816)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The patches this recipe refers to don't exist so nobody can have
been using this so lets just remove it.
(From OE-Core rev: 29916bfbbea3bb785e8e4d332848b19ebc0a3ea9)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Default ABI for mips64 is n64 so make it default in binutils
(From OE-Core rev: 99f859612e16d636702f9bb11825259d90e468f2)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
WARNING: For recipe binutils, the following files/directories were installed but not shipped in any package:
WARNING: /usr/bin/embedspu
(From OE-Core rev: 447d77b9e33a68f748c96a152eae19148e43ed7c)
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
A comment in the binutils sources indicate that it should support all of the
-march= parameters that gcc supports. The tune validations noted that
gas failed on -march=armv5e.
It is not yet clear to me if this patch belongs upstream or not.
(From OE-Core rev: 4f2503ebdb5c491b758cbdf6eb7df96d4295c24f)
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Dont use autotools, it really not so autoconf like.
the configure script gets updated with every release of zlib
and we overwrite that. Instead use the upstream provided
configure
copyright year was changed in zlib.h which caused change in
LIC_FILE_CHECKSUM
fix.inverted.LFS.logic.patch is already applied upstream so drop it
Drop the configure.ac and Makefile.am scripts since we do not
autoreconf anymore and do not inherit autotools anymore
Bump PR for depending recipes so a rebuild it ensues so that
they dont depend on .la anymore
and add missing dependencies discovered during incremental
build
(From OE-Core rev: 50ad5230ea9e0982cdfda23fb9fcfccf89d28f29)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
In order to actually generate a toolchain (with
bitbake meta-toolchain) that supports gold, binutils-
cross-canadian also needs to be built this way.
(From OE-Core rev: 0eb3a0f3ee857a17d80598ea2f0b1f3d14ac6d30)
Signed-off-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
x86_64 opensuse 11.4 has bintuils version 2.21, and when
binutils_2.21 recipe is built for x86_64 target then, the invocation
of distro gcc fails with errors like this:
/usr/lib64/gcc/x86_64-suse-linux/4.5/../../../../x86_64-suse-linux/bin/as:
symbol lookup error: /usr/lib64/gcc/x86_64-suse-linux/4.5/..
make[2]: *** [sysinfo.o] Error 1
The issue rootcaused as incompatible LD_LIBRARY_PATH while running
the distro gcc.
As per Martin Jansa gentoo also sees similar issue with binutils 2.22
recipe.
This commit fixes the issue by clearing the LD_LIBRARY_PATH for
distro gcc (CC_FOR_BUILD)
This Fixes bug: [YOCTO #1833]
(From OE-Core rev: be769d948a9c26d55d98fbf38fc0b109edb40c3f)
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Right now for cross recipes e.g. gcc-cross and binutils-cross
we specify --disable-nls .... --enable-nls on configure cmdline
the --enable-nls coming from gettext bbclass.
So we disable nls for all cross inheriting recipes in gettext
bbclass and then we remove the extra --disable-nls in gcc-cross
and binutils-cross
This patch needs testing. Please help
(From OE-Core rev: d66b379f809b9c75981848fcc71ed5de13382bf7)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Khem Raj
Phil Blundell
Richard Purdie
Saul Wold
Zhenhua Luo
Marcin Juszkiewicz
Laurentiu Palcu
Enrico Scholz
Ross Burton
Robert Yang
Matthew McClintock
Andy Ross
Mark Hatle
Martin Jansa
Christopher Larson
Derek Buitenhuis
Nitin A Kamble