| Age | Commit message (Collapse) | Author | Files |
|---|
|
We want to revert to default gcc behavior to support oe-core's ability
to change the libdir.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
It can be alarming to attempt to exclude GPLv3 from an
image but find that libstdc++ and libgcc still show it.
We indicate the license for each package to show libraries
that really are just GCC-3.0-with-GCC-exception.
Signed-off-by: Joe Slater <jslater@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
Whilst gcc doesn't have any source to fetch, it still needs a fetch task so that
a world fetch can run without errors. So instead of deleting the fetch task,
stub it.
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
The current implementation of shared work for gcc is at best confusing. It relies
on the fetch/unpack/patch tasks having exactly the same stamps and if this gets
broken for some reason, its hard to figure out what the problem is. It also
leads to complex code in bitbake.
The benefits of shared work for gcc are clear but a better approach is needed. This
patch adjusts things so that a single new recipe (gcc-source) provides the
fetch/unpack/patch/preconfigure tasks, the rest of gcc simply depends on these tasks
and have no fetch/unpack/patch tasks of their own.
This means we should get the significant benefits (disk usage/performance) of the
single source tree but in a way which has less potential for problems and is
easier for people to understand. The cost is an extra recipe/some inc files
which is probably a good tradeoff.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
I disabled this patch as it became obsolete some time ago but forgot to
remove it, this cleans things up.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
While compiling gcc-crosssdk-initial-x86_64 on some host, there is
occasionally failure that test the existance of default.h doesn't
work.
...
| tmp/work-shared/gcc-4.9.1-r0/gcc-4.9.1/gcc/calls.c:1240:
error: 'STACK_CHECK_MAX_VAR_SIZE' was not declared in this scope
...
The reason is tm_include_list='** defaults.h' rather than
tm_include_list='** ./defaults.h'
So we add the test condition for this situation.
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
If configure or any of the components it uses from the shared work directory
change, do_configure may fail.
An existing do_preconfigure was created to handle these conditions, but
a 'sed' operation was missed, and a call to gnu-configize was also missed.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The first patch fixes the ICE in dwarf2out_var_location, at
dwarf2out.c.
r212171:
* except.c (emit_note_eh_region_end): New helper function.
(convert_to_eh_region_ranges): Use emit_note_eh_region_end to
emit EH_REGION_END note.
* jump.c (cleanup_barriers): Do not split a call and its
corresponding CALL_ARG_LOCATION note.
But it introduced a regression issue:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63348
so backport the fix for the regression as well:
r215613:
PR rtl-optimization/63348
* emit-rtl.c (try_split): Do not emit extra barrier.
Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
Various pieces of the code assume that the --sysroot option gets passed
into the compiler tools. By having a "sane" default, we don't always
spot when this occurs and this can later show up as breakage in sstate,
or in usage of the external toolchain.
We've long since talked about poisoning the default such that it will
break unless the correct option is specified. This patch does just that.
If this patch causes something to fail to build, it most likely means
the various compiler flags and commands are not correctly being passed
through to the underlying piece of software and that there is a real
problem that needs fixing, its not the fault of this patch.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This fixes gcc bug 6144, which in my case exhibited itself as a kernel
module that failed to load. This was because static platform_data
structures were being corrupted with the optimiser being set to any
value other than -O0.
Originally-submitted-by: Peter Urbanec <openembedded-devel@urbanec.net>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Some architectures can mix different TARGET_OS values, in most cases
we just use one but in the ppc case, can use two different values. In this
case, to use one toolchain with both, we need to ensure the symlinks exist.
This isn't ideal but does fix the ppc toolchains for the release, after
which better ways of handling this can be investiaged. Without this, failures
in the C++ toolchain are seen.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
There is a race where:
NOTE: recipe libgcc-initial-4.9.1-r0: task do_configure: Started
NOTE: recipe gcc-runtime-4.9.1-r0: task do_preconfigure: Started
| checking build system type... /home/pokybuild/yocto-autobuilder/yocto-worker/nightly-deb/build/build/tmp/work-shared/gcc-4.9.1-r0/gcc-4.9.1/libgcc/../config.sub: line 1711: syntax error near unexpected token `;;'
| /home/pokybuild/yocto-autobuilder/yocto-worker/nightly-deb/build/build/tmp/work-shared/gcc-4.9.1-r0/gcc-4.9.1/libgcc/../config.sub: line 1711: ` ;;'
| configure: error: /bin/bash /home/pokybuild/yocto-autobuilder/yocto-worker/nightly-deb/build/build/tmp/work-shared/gcc-4.9.1-r0/gcc-4.9.1/libgcc/../config.sub x86_64-linux failed
| WARNING: exit code 2 from a shell command.
so we need to make sure the preconfigure task executes in all shared
work contexts.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
change use of eglibc related variabled to glibc equivalents
Signed-off-by: Khem Raj <raj.khem@gmail.com>
|
|
musl e.g. is configured to not use fixed-include
which is an improvement btw. but libgcc-initial configure
has tests which probe for limits.h and since we put
it in include-fixed/ dir and that dir does not appear
in gcc's internal default search path the configure tests
for CPP detection fail and libgcc-initial can not be compiled.
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The gcc-runtime recipe builds the gcc libraries including libstdc++ with
$TARGET_CC_ARCH flags, which include -march=FOO flags that affect
whether atomic instructions are available. This causes an ABI
incompatibility when the compiler by default generates code for less
capable architectures. For example, gcc-runtime libraries on a
Cortex-A8 are built with a different C++11/C++14 mutex implementation
than is used code compiled outside OE and without architecture-specific
flags.
This commit fixes the problem specifically for ABI issues related to
atomic instructions available in ARMV6 and subsequent architectures.
Other ABI incompatibilities may remain in other architectures.
See: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62100
Signed-off-by: Peter A. Bigot <pab@pabigot.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
A long-standing bug in gcc turns out to cause problems with unpatched
Linux versions due to improved optimization enabled by gcc 4.9. The
upstream fix missed the gcc-4.9.1 cut-off. It's also been applied
upstream to the 4.8 branch so is being added for OE's 4.8 as well.
Signed-off-by: Peter A. Bigot <pab@pabigot.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
musl does not support IBM 128 long double for ppc, instead of
doing complex overrides move it into a pythong snippet which
is easier to read and more compact.
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This variable is used to ensure the proper version of --with-float=FOO
is passed to gcc's configure script. gcc also has a --with-fpu=FOO
option that means something different. To avoid confusion, change the
names to be consistent.
Signed-off-by: Peter A. Bigot <pab@pabigot.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Peter A. Bigot <pab@pabigot.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
--enable-libunwind-exceptions was removed from gcc at release 3.4.3
about ten years ago.
Signed-off-by: Peter A. Bigot <pab@pabigot.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Apply to gcc 4.9 the recent fix to the --with-gxx-include-dir override.
Original OE-Core rev: 5a2ff3e8f7cd7a47a5ab4e581847ecc4df87fca
Signed-off-by: Peter A. Bigot <pab@pabigot.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
0037-gcc-4.8-PR56797.patch was originally added as an OE backport during
4.8.0. Upstream merged it in 4.8.1, and it was present in 4.9.0.
The original patch still applies to 4.9.1 (and presumably 4.8.2), but
now is modifying store_multiple_sequence instead of
load_multiple_sequence (the two functions are nearly identical). It may
or may not be necessary in store_multiple_sequence, but absent a bug
report upstream supporting its application in this case, or a least an
updated comment and upstream status in the patch, I think this patch
should be dropped.
Signed-off-by: Peter A. Bigot <pab@pabigot.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Consistent use of whitespace in multi-line assignment, with special
focus on OECONF modifications. Quotes on separate lines, four-space
indentation, one value per line.
Signed-off-by: Peter A. Bigot <pab@pabigot.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
We already indicate our intentions to use ld.bfd by
specifying it in configure using --with-ld which works
ok unless here where we manually create symlinks to
binutils-cross components, when we use ld-is-gold feature
default ld points to gold and this symlinking has to be
aware of the fact that we configured binutils and gcc-cross to use
gold as default ld but gcc-cross-initial uses BFD ld
This would be visible when using gold and rebuilding
eglibc
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Fix an issue on a multilib configuration that contains more then 1 multilib.
I.e. on MIPS64:
DEFAULTTUNE = "mips64"
MULTILIBS = "lib32n:mips64_n32 lib32:mips32"
While normally you'd use 'libn32', the above is legal.
With the startswith code, the system will look to expand the 'lib32' element
and find the 'lib32n' instead, and will result in a warning:
lib32 doesn't have a corresponding tune. Skipping...
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
|
|
Drop patches which are already available in 4.9.1
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
|
|
The gcc-ar.o, gcc-nm.o, gcc-ranlib.o and errors.o included
config.h which was a generated file. But no explicity rule
to clarify the dependency. There was potential building
failure while parallel make.
For gcc-ar.o, gcc-nm.o and gcc-ranlib.o, they were compiled from one C
source file gcc-ar.c, we add them to ALL_HOST_BACKEND_OBJS, so the
'$(ALL_HOST_OBJS) : | $(generated_files)' rule could work for these
objects.
For errors.o, it is part of gengtype, and the gengtype generator program
is special: Two versions are built. One is for the build machine, and one
is for the host. We refered what gengtype-parse.o did (which also is part
of gengtype).
[YOCTO #6568]
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
|
|
In subdir 'gcc', Most C source files included config.h which was
generated by a rule. But no related prerequisites was added to
the C source compiling rule. There was potential building failure
while makefile enabled parallel.
The C source compiling rule used suffix rule '.c.o', but the suffix
rule doesn't support prerequisites.
https://www.gnu.org/software/make/manual/html_node/Suffix-Rules.html
We used the pattern rule '%.o : %.c' to instead, and add the config.h
as its prerequisite
We also moved the '%.o : %.c' rule down to the 'build/%.o :' rule, which
makes '%.o : %.c' rule doesn't override 'build/%.o :'.
[YOCTO #6568]
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
MULTILIB_OPTIONS takes the parameters which trigger a given multilib to be
selected. It supports *one* option per multilib, '/' separated. Spaces
separate options used to generate additional multilib combinations.
Adding in all of CFLAGS to this is therefore clearly a really bad idea
but how do we fix things?
The best option I've come up with so far is a list of whitelist variables
to use to trigger the multilibs. Its populated with the standard multilibs
we support, anyone setting up an advanced multilib can populate the variable
with the correct trigger parameters.
This has the advantage of simplifying the code and allowing us to remove
the code filtering blocks since there is no longer option duplication. Testing
after this change shows a much improved sdk toolchain functionality.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
* GLIBC_DYNAMIC_LINKER64 reglex does not work for rs6000/linux64.h,
update it.
* it turns out that UCLIBC_DYNAMIC_LINKER reglex will strip the 32/64
chars from UCLIBC_DYNAMIC_LINKER64/UCLIBC_DYNAMIC_LINKER32, add '\b'.
my two PCs: Centos 6.5 (python 2.7.5) and Fedora 13 (python 2.7.3)
Signed-off-by: Ting Liu <ting.liu@freescale.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
It was observed that code using STLport 4.6 fails to compile under the
SDK with the following error message:
.../includes/cstddef:38:46: fatal error: ../4.7.2/cstddef: No such file
or directory
STLport 4.6 (screwily) assumes that the C++ system headers live in a
gcc-versioned subdirectory, for gcc>=3.0; cf
http://sourceforge.net/p/stlport/code/ci/STLport-4.6-patch/tree/stlport/config/stl_gcc.h#l269.
This assumption is *almost always* valid, because that matches the
default setting of --with-gxx-include-dir. We can match that behavior by
appending "/${BINV}" to our own --with-gxx-include-dir settings.
Natinst-CAR-ID: 446449
Natinst-Reviewboard-ID: 57209
Acked-by: Ken Sharp <ken.sharp@ni.com>
Acked-by: Ben Shelton <ben.shelton@ni.com>
Signed-off-by: Richard Tollerton <rich.tollerton@ni.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
|
|
When enabling a lib32-gcc in a 64 bit build, without doing any
other configuration, the mutilib dir is unspecified, which is
represented internally in gcc as "." and as such uncovers an
invalid free on a non-malloc'd pointer.
As suggested by the gcc folks, simply make sure the "." case
is also stored in a malloc'd pointer, so that the intended
runtime behaviour of the code remains unchanged.
Patch has been accepted by upstream maintainers of gcc.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
While we're not going to package the libgcc component as part of the SDK,
we do need to generate it to get the unwind, and quadmath headers. Without
this change it is not possible to build eglibc or other components that
require these headers with the SDK toolchain.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The gengtype patch we apply to gcc aims to ensure that the build and host
config headers don't get confused. We're seeing build failures where
both headers have been included, likely due to a race over the configuration
files.
It seems the gengtype-lex.c file isn't being regenerated when it should
and the unconditional inclusion of bconfig.h is resulting in these issues.
The fix is therefore to remove the file, forcing its regeneration.
[YOCTO #6393]
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The do_configure_prepend was duplicated in gcc-4.X.inc and
gcc-configure-common.inc leading to confusion when reading the resulting
do_configure task where the file was processed twice.
The only difference was the removal of the include line for gcc 4.8/4.9.
On mingw were were seeing two issues, firstly that the if statements meant
the values we wanted weren't being set, the second that the include
paths were still wrong as there was no header path set.
To fix the first issue, the #ifdef conditionals were removed, we want
to set these things unconditionally. The second issue is addressed by
setting the NATIVE_SYSTEM_HEADER_DIR variable here (it was already
set in t-oe).
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
[OE-core bug #6270] - https://bugzilla.yoctoproject.org/show_bug.cgi?id=6270
Signed-off-by: Alexandru-Cezar Sardan <alexandru.sardan-KZfg59tc24xl57MIdRCFDg@public.gmane.org>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Fixes [YOCTO #4497]
Usage of FILESPATH is discouraged, since it can make recipes harder to
bbappend. Instead FILESEXTRAPATHS should be used to extend the path.
Signed-off-by: Petter Mabäcker <petter@technux.se>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
We need to handle the UCLIBC_* linker variables in the same way
as we do the GLIBC_* ones to allow uclibc multilib to work properly.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
(From OE-Core rev: f051216ea373f166016b15bbd2a2a6f136430372)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
There is a fix about to go into bitbake to ensure that datastores
being accessed with a name other than "d" are correctly reflected
in checksums. This will cause this function to add in a number of
dependencies we don't want.
These do need to be properly unravelled in due course but would
only really affect multilib builds. For now therefore just exclude
the variables as per the old behaviour.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The class itself currently does nothing. The idea is to mark all recipes that
make use of the texinfo utilities. In the future, this class could be used to
suppress the generation/formatting of documentation for performance,
explicitly track dependencies on these utilities, and eliminate Yocto's
current dependency on the host system's texinfo utilities.
Signed-off-by: Max Eliaser <max.eliaser@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Within the OE build environment, we supply the correct fpu settings. These
only need to be spelt out for the on-target gcc.
Doing this means the checksums for the core compiler don't depend on the fpu
settings. We exclude the compiler tunes for similar reasons, it doesn't need
to influence the compiler build.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Since we no longer build target libs within gcc-cross, we can drop the
TARGET_CC_ARCH flags and hence make it independent of tune.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
As far as I can tell this variable is now completely unneeded. It would
only ever get used in target builds and these are now correctly done
in the target environment namespace, not any of our cross environments.
As such, CC and other variables contain the correct compilers and other
tune options and these are correctly picked up when building libgcc,
libstdc++ and others.
I tried to figure out where else these would make any sense and couldn't
find anything. Builds appear fine without them so lets drop the complexity
including the patch adding in this flag to gcc.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This command modifies ${S} and can race against other tasks running do_configure and
having the scripts disappear from under them. To avoid this move to its own
task and work on the shared work directory as a common task.
It needs to be a python task to avoid lots of shell exported variables as
dependencies.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Currently the gcc builds are building copies of the target libraries
that we never use (it isn't installed in do_install). This is a rather
pointless waste of cpu time.
Instead just compile the host targets. Comparing the package output of
this compared to a previous build shows that the unwind.h header is
missing since its provided by gcc. Fix this simply by copying it in.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This allows them to co-exist together in the native sysroot, with one
set of cross tools per target architecture.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The base_contains is kept as a compatibility method and we ought to
not use it in OE-Core so we can remove it from base metadata in
future.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
