Now that KERNEL_LD uses ld.bfd, we need to ensure that it exists when using an
external toolchain.
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
* mel/master:
external-sourcery-toolchain: kill more bits we don't want in eglibc-utils
Work around upstream FILESDIR/FILESPATH changes
external-sourcery-toolchain: fix dev vs staticdev packaging issue
external-sourcery-toolchain: kill oprofile bits
external-sourcery-toolchain: silence GNU_HASH warning for libquadmath
tcmode: prefer a matching gcc version for target recipes
external-sourcery-toolchain: add libquadmath packages
external-sourcery-toolchain: kill ${exec_prefix}/lib/locale in do_install_locale
external-sourcery-toolchain: also package ${prefix}/libexec
external-sourcery-toolchain: forgot to rstrip the trailing newline
Conflicts:
recipes/meta/external-sourcery-toolchain.bb
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
As do_install_locale already copied these into the sysroot to be picked up by
eglibc-locale, there's no reason to keep the files around for ourselves.
Note: this should go into the oe-core eglibc-package.inc.
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
As do_install_locale already copied these into the sysroot to be picked up by
eglibc-locale, there's no reason to keep the files around for ourselves.
Note: this should go into the oe-core eglibc-package.inc.
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
This is needed, as even if libexecdir isn't ${prefix}/libexec, our already
built eglibc will be expecting its files there. Without this, we get package
QA failures with current oe-core, which changed libexecdir to ${libdir}/${PN}.
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
* gh-public/danny:
external-sourcery-toolchain: ship the multilib link in the -dev package
external-sourcery-toolchain: put linux headers in eglibc packages
toolchain-scripts: add TOOLCHAIN_PATH_ADD to PATH
tcmode: use the class override for PATH add
tcmode: don't ship a toolchain in the sdk for now
When building glibc, this is what usually happens -- the headers are
sanitized, then they're included in the libc development package(s), so mirror
that here. Include the headers with eglibc-dev rather than splitting out
separate linux-libc-headers packages.
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Long term, may want to add the ability to ship the cross-canadian content from
the external toolchain.
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
To avoid issues when a user installed a Sourcery toolchain to the default path
and let it modify their PATH, we iterate over PATH and remove the problematic
elements. Unfortunately, in doing so, we inadvertently forced an early
expansion of PATH due to getting it in its expanded form, then setting PATH
based on that form. Instead, operate against the unexpanded PATH.
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
The toolchain installer can end up modifying the user's PATH via their
dotfiles, but if the ia32 sourcery g++ toolchain is always in the user's path,
things can fail to build, as the ia32 toolchain provides non-prefixed binaries
(e.g. 'gcc', 'ld').
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
THis avoids eglibc being rebuilt when trying to include packages like
eglibc-pcprofile. This includes world, and the lsb packagegroup.
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
* some stuff related to oprofile was being built but it was not being shipped,
added that to be packged now.
Signed-off-by: Fahad Usman <fahad_usman@mentor.com>
This change allows to generate libgomp packages for external toolchains
with libgomp support.
Signed-off-by: Vladimir Zapolskiy <vladimir_zapolskiy@mentor.com>
Note from kergoth: this is a temporary build fix, likely not an ideal fix for
the underlying problem.
Signed-off-by: Muhammad Shakeel <muhammad_shakeel@mentor.com>
BitBake doesn't run event handlers in registration order, which means there's
no way to know if a given bit of code (e.g. the tcmode set of
EXTERNAL_TOOLCHAIN_SYSROOT) has been set, from a different event handler. So
go back to using ${@}.
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
