If we used -fuse-ld gcc option, then it does not work
ending in
collect2: fatal error: cannot find 'ld' compilation terminated.
This is because we are not creating proper symlinks for BFD
and Gold linker in gcc installation
Secondly, we end up with dangling fortran compiler symlinks
if fortran is not enabled when confguring gcc, therefore
create these symlinks only when fortran support is enabled
in gcc
(From OE-Core rev: f84614c082406287e608860741a82edd2f3c9bca)
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>
The path to where to install and find the sysroot components is used
in many places. This warrants it to get its own variable.
(From OE-Core rev: 70a84b525470f72339568409daf84845904e4cab)
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.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>
Its sad we need to do this but do_install and do_populate_sysroot appear
to change the files in ${B} and this breaks if something works in parallel
like gcc_stash_builddir. We've seen a few too many race errors on the
autobuilder which appear to be from this so make things run in sequence
deterministically for now.
An example failure was this from do_populate_sysroot whilst stash_builddir was running
in parallel:
ERROR: gcc-cross-initial-arm-6.3.0-r0 do_populate_sysroot: split_and_strip_files: 'file
/home/pokybuild/yocto-autobuilder/yocto-worker/nightly-arm-lsb/build/build/tmp/work/x86_64-linux/
gcc-cross-initial-arm/6.3.0-r0/sysroot-destdir/home/pokybuild/yocto-autobuilder/yocto-worker/
nightly-arm-lsb/build/build/tmp/work/x86_64-linux/gcc-cross-initial-arm/6.3.0-r0/recipe-sysroot-native/
usr/bin/arm-poky-linux-gnueabi.gcc-cross-initial-arm/arm-poky-linux-gnueabi-gcov' failed
Exception: subprocess.CalledProcessError: Command '('patchelf-uninative', '--set-interpreter',
'/home/pokybuild/yocto-autobuilder/yocto-worker/nightly-arm-lsb/build/build/tmp/sysroots-uninative/
x86_64-linux/lib/ld-linux-x86-64.so.2', '/home/pokybuild/yocto-autobuilder/yocto-worker/nightly-arm-lsb/
build/build/tmp/work/x86_64-linux/gcc-cross-initial-arm/6.3.0-r0/sstate-build-populate_sysroot/
recipe-sysroot-native/usr/bin/arm-poky-linux-gnueabi.gcc-cross-initial-arm/arm-poky-linux-gnueabi-gcov')' returned non-zero exit status 1
Subprocess output:
missing section headers
(From OE-Core rev: 75b76eb08c7e344142f0326605aeb6b24e61e38e)
(From OE-Core rev: 7265e99d0b4b662ca7f23d480f2845a309ac57fc)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
If we don't do this, things break with rm_work which removes things
before the task completes causing task failures.
(From OE-Core rev: bfea0e6a5aee9ce15bf20bcb2781329e6281284c)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When we stashed the gcc build directory for use in generating the various runtimes
we were being lazy and just used the staging directory. With recipe specific
sysroots this means we're copying a large chunk of data around with the cross
compiler which we don't really need in most cases.
Separate out the data into its own task and inject this into the configure
step. We have to do that here since autotools will wipe out ${B} if it thinks
we're rebuilding and we therefore have to time its recreation after that.
This also takes the opportunity to remove some pointless (as far as I can tell)
conditionals from the do_install code.
(From OE-Core rev: dcf15ccf3cc9d55e77228ba8d526f967fc9791b4)
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>
gcc-cross target recipes should not depend on SDK_SYS but started to
after recent changes. Remove the dependency to stop this (its caused
by shared code in do_install). The compiler names contain SDK_SYS
so changes would be correctly handled via other means.
(From OE-Core rev: 2b5761350a074de2e1a6db19621945fba39089fc)
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>
We explicitly set the hash style to gnu in our LDFLAGS. Setting the default to
this in the toolchain, while convenient, actually hides bugs, as a failure to
obey LDFLAGS isn't noticed. By removing this, it's not dissimilar to how we
poison the sysroot -- rather than relying on the default, notice right away if
somoeone isn't obeying the needed flags.
This will result in a failure to obey LDFLAGS causing a GNU_HASH QA failure,
which is what's often seen with external toolchains. This brings us all on the
same page, and makes sure a failure to obey LDFLAGS is seen early.
This is limited to cross, to retain ease of use for SDKs.
(From OE-Core rev: fa436aeb3242cbfdbbe16d448d45bce8eb5b74fd)
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
PATCH_GET is no longer exported by bitbake.conf, so no longer needs
to be pruned from the gcc-cross do_compile() environment.
(From OE-Core rev: 386d1617f1e6e293506b751faa41ff8c37348973)
Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Modifications to to support build for TCLIBC="baremetal"
(From OE-Core rev: 9e7da158e0e7628c22013fbee34270f7d2db621f)
(From OE-Core rev: 92eb54c6b7424770f96a8e14e792cbe5124968c3)
Signed-off-by: Juro Bystricky <juro.bystricky@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
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.
(From OE-Core rev: 21c2cfff14442cf224e3568bdbb9bcd4070be247)
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
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.
(From OE-Core rev: 04b725511a505c582a3abdf63d096967f0320779)
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.
(From OE-Core rev: 285d3579727177e6962d7ad16677429e7dec65f4)
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.
(From OE-Core rev: d971db8b2259e4c35b871cccf130fba193849560)
Signed-off-by: Peter A. Bigot <pab@pabigot.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>
(From OE-Core rev: 5a2ff3e8f7cd7a47a5ab4e581847ecc4df87fca3)
Signed-off-by: Richard Tollerton <rich.tollerton@ni.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
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.
(From OE-Core rev: 74d8866814aec520822518cc4cb8a942f7069bf7)
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.
(From OE-Core rev: a2c5509520d5c3e082f55844e6545d0309565f8f)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Rather than building the whole of libgcc to obtain the unwind.h header
file, simply configure it and then install the file. This avoids copying
chunks of data around when we don't need to and building the same thing
twice.
After doing this we need to make sure the target build directory exists
in the libgcc case since it will no longer be created automatically.
(From OE-Core rev: 459e4dc25462771038459567c22e87d4cd38b117)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Instead of using oe.path.relative, use the Python Standard Library function
os.path.relpath.
(From OE-Core rev: 90c3a0401c566e26d89a5c0410b2a51fe27b95b2)
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
gcc 4.8 fortran presents some challenges:
* libquadmath headers need to be in the libexec include dir. It turns out
to be easiest just to manually do this.
* libgfortran configure needs libquadmath to be compiled. This means
a separate recipe is needed (the alternative is gross hacks)
* the libtool uses to link libgfortran doesn't have our improved rpath
handling and puts bogus RPATHS into the libraries. We can avoid this
by tweaking libtool with sed.
This patch resolves those issues. Any user of fortran does need to DEPEND
on libgfortran in order to trigger it to build but this shouldn't be a major
issue.
(From OE-Core rev: a5e7ee5770b9e0cf719c573efffd874440f74289)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* Remove the duplicate EXTRA_OECONF_PATHS that is overwritten
* Merge the do_compile and do_compile_prepend
* Group dependency and configuration variables together
(From OE-Core rev: 430b1d6fcbf60df35036fe5bbf8e55f7fb8f5341)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The include was useful historically, its not anymore so lets remove it.
This should have no functional change except on any layers directly depending
on it or gcc-cross.inc but even then it would only impact sh4 and is easily
fixed if there was a problem.
(From OE-Core rev: e7e8fe11c34bf05179f3bbaa2fb1af7b7125696a)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
gcc-cross cannot build without linux-libc-headers but doesn't explicitly depend on
it relying on the implied dependency through libc. With cases where pieces
can be installed through sstate, we now need this explicit dependency to
ensure builds with partial sstate work.
(From OE-Core rev: 65e5670ef429bb6c348decb1804e425f1c4d7c61)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
shift the cmdline argument after determining
there is atleast 1 argument to shift otherwise
it ends up in error when no argument is specified
(From OE-Core rev: 4cd306ab0c9404a31f38c7e5fb67fe32fdca3d4b)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
dash does not like >& so be explicit and say 2>&1
(From OE-Core rev: 1d262630853e65be9167d904b934b581acf64182)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Dont use -q to grep we pipe to /dev/null anyway all we
care is the return status of grep
(From OE-Core rev: 9c575b5f37ee53eaf26bfdc468d4cbb661db703f)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
gcc-cross is installed into a package architecture specific directory
and is not meant to be machine specific. This patch replaces MACHINE_ARCH
with PACKAGE_ARCH to ensure this is really the case.
This was found by examining sstate checksums.
(From OE-Core rev: b3ca57418f47b16ed0a63d3291bf7d31e3ca5a45)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This script will be generated into the build directory of gcc-cross
It should be testing gcc and g++. libstdc++ tests are not run since
we build them as part of gcc-runtime but we can test them here by
building them with 'make all' and then running the tests
The script expects passwordless ssh access to target and is used
in form
./arm-oe-linux-gnueabi-testgcc kraj@192.168.7.2
inside the builddir of gcc-cross
(From OE-Core rev: 130b534fdb0b292158981a12d7d5f01c1a14cb2a)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Having one monolithic packages directory makes it hard to find things
and is generally overwhelming. This commit splits it into several
logical sections roughly based on function, recipes.txt gives more
information about the classifications used.
The opportunity is also used to switch from "packages" to "recipes"
as used in OpenEmbedded as the term "packages" can be confusing to
people and has many different meanings.
Not all recipes have been classified yet, this is just a first pass
at separating things out. Some packages are moved to meta-extras as
they're no longer actively used or maintained.
Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
