When using sstate, or performing an incremental build any change to
the do_stash_locale() will cause a build failure because
do_stash_locale() was destroying the results obtained from the
do_install() with several mv operations. A recent change to
do_stash_locale() for a different problem illustrated a number of
build failures for users in the community.
To fix the problem, do_stash_locale() must use copy operations instead
of the mv operations. Because this is changed to a copy, the sysroot
and package stage need to remove the files that would have been
previously removed. The correct "fixup" code to deal with the removal
already existed in the previous do_poststash_install_cleanup(). All
that needed change was the path to where to remove the files
from the sysroot and package stages.
In order to force a re-compilation of glibc some unused white space
was removed from do_compile() for glibc. I could not find any other
way around this and we don't want to have all the community folks to
have another iteration where they have to remove their tmp directories
or purge some portion of the sstate. It also makes this change
bisectable. If the change to the glibc is not included, it will fail
with the following message:
=====
| DEBUG: Executing shell function do_prep_locale_tree
| tar: i18n: Cannot stat: No such file or directory
| tar: Exiting with failure status due to previous errors
| gzip: /poky/build/tmp/work/core2-64-poky-linux/glibc-locale/2.29-r0/locale-tree//usr/share/i18n/charmaps/*gz.gz: No such file or directory
=====
After this one time change I tested changing only the
do_stash_locale() function and it now works well because it is
deterministically operating off the sstate data or a local build.
(From OE-Core rev: fedc57a41a15bca1d96d14e25e2df0bb1eca904d)
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The do_stash_locale was not working consistently across the 4 build
configurations and the multilib, usrmerge configuration would fail
entirely with the obscure message:
| DEBUG: Executing shell function do_prep_locale_tree
| tar: i18n: Cannot stat: No such file or directory
| tar: Exiting with failure status due to previous errors
| gzip: /poky/build/tmp/work/core2-64-poky-linux/glibc-locale/2.29-r0/locale-tree//usr/share/i18n/charmaps/*gz.gz: No such file or directory
| WARNING: /poky/build/tmp/work/core2-64-poky-linux/glibc-locale/2.29-r0/temp/run.do_prep_locale_tree.124690:1 exit 1 from 'gunzip $i'
Here is the 4 build configurations without the patch applied:
A) x86-64 no multilibs, no usrmerge
find ./tmp/work/*/glibc/2.29-r0/stashed-locale -type f |grep -v nscd.service |wc -l
909
B) x86-64 no multilibs, usrmerge
find ./tmp/work/*/glibc/2.29-r0/stashed-locale -type f |grep -v nscd.service |wc -l
909
C) x86-64 multilibs, no usrmerge
find ./tmp/work/*/glibc/2.29-r0/stashed-locale -type f |grep -v nscd.service |wc -l
885
D) x86-64 multilibs, usrmerge
find ./tmp/work/*/glibc/2.29-r0/stashed-locale -type f |grep -v nscd.service |wc -l
864
The issue here is that all the moves should be processed first, then a
copy should be made of the lib directories, but only in the case they
are different when using the usrmerge feature. Even though the build
worked for the multilib configuration without usrmerge, the content
was not the same.
After applying the patch the same number of files are in all the
configurations. The list of files was also diffed, after normalizing
the directory names to ensure all the correct files were copied.
Ultimately there are probably additional files that should be pruned
from what is copied to the stated_locale, but the purpose of this
patch is make it 100% consistent between the build types and fix the
builds.
(From OE-Core rev: 33c2e7b4944af22ca47b53d1f85d03426f169bb7)
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The build of glibc fails when you have multilibs enabled + the distro
feature usrmerge. Here is an example configuration:
===
MACHINE = "qemux86-64"
VIRTUAL-RUNTIME_init_manager = "systemd"
DISTRO_FEATURES_append = " systemd "
DISTRO_FEATURES_append += " usrmerge"
require conf/multilib.conf
MULTILIBS = "multilib:lib32"
DEFAULTTUNE_virtclass-multilib-lib32 = "x86"
===
This will fail with the following error:
NOTE: Executing SetScene Tasks
NOTE: Executing RunQueue Tasks
ERROR: glibc-2.28-r0 do_poststash_install_cleanup: Function failed: do_poststash_install_cleanup (log file is located at /poky/build/tmp/work/core2-64-poky-linux/glibc/2.28-r0/temp/log.do_poststash_install_cleanup.107893)
ERROR: Logfile of failure stored in: /poky/build/tmp/work/core2-64-poky-linux/glibc/2.28-r0/temp/log.do_poststash_install_cleanup.107893
The fix is to not perform the rmdir check when using the multilib + usr/merge, namely:
if [ "${libdir}" != "${exec_prefix}/lib" ] && [ "${root_prefix}/lib" != "${exec_prefix}/lib" ]; then
This will evaluate as follows (collecting the output from bitbake -e glibc)
* no multilibs no usrmerge
if [ "/usr/lib" != "/usr/lib" ] && [ "/lib" != "/usr/lib" ]; then
* no multilibs yes usrmerge
if [ "/usr/lib" != "/usr/lib" ] && [ "/usr/lib" != "/usr/lib" ]; then
* yes multilibs no usrmerge
if [ "/usr/lib64" != "/usr/lib" ] && [ "/lib" != "/usr/lib" ]; then
* yes multilibs yes user merge
if [ "/usr/lib64" != "/usr/lib" ] && [ "/usr/lib" != "/usr/lib" ]; then
(From OE-Core rev: c5640f8c8663c8f81125bf7c5bc2ef8e9fe55315)
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
fix below error:
file /usr/include/bits/procfs-id.h conflicts between attempted installs of lib32-libc6-dev-2.29-r0.armv7vet2hf_vfp and libc6-dev-2.29-r0.aarch64
file /usr/include/bits/procfs.h conflicts between attempted installs of lib32-libc6-dev-2.29-r0.armv7vet2hf_vfp and libc6-dev-2.29-r0.aarch64
file /usr/include/bits/shmlba.h conflicts between attempted installs of lib32-libc6-dev-2.29-r0.armv7vet2hf_vfp and libc6-dev-2.29-r0.aarch64
(From OE-Core rev: 1e9120096da81171e9213b0b78df0aff7002de15)
Signed-off-by: Changqing Li <changqing.li@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The AArch64 little-endian ABI requires that the dynamic loader is always
available at /lib/ld-linux-aarch64.so.1. Similarly, the big-endian ABI
requires that the dynamic loader is always available at
/lib/ld-linux-aarch64_be.so.1.
glibc-package.inc contains code that tries to ensure this, but
unfortunately it is defeated by the combination of multilib and usrmerge
because it does not take into account that /lib is the same as /usr/lib
with usrmerge when it adds the loader path to libc_baselibs and when it
attempts to show that /usr/lib is empty in do_poststash_install_cleanup.
This results in the symlink not being included in the package and a build
failure due to rmdir failing.
Richard Purdie also suggested[1] that ${nonarch_base_libdir} should not be
used as a synonym for /lib in this case.
This hopefully-fixed version always sets ARCH_DYNAMIC_LOADER and then uses
${root_prefix}/lib/${ARCH_DYNAMIC_LOADER} to refer to the dynamic loader
which works with both multilib and usrmerge. Since ARCH_DYNAMIC_LOADER is
only non-empty if the symlink is required, the code to create it can move
to do_install_append. Then do_poststash_install_cleanup needs to be taught
that ${exec_prefix}/lib may not be empty if the dynamic loader symlink is
there.
It appears not to be possible to specify the name of the loader via a
variable with an override, since the _aarch64 override is applied even for
_aarch64-be, so I've set the loader name using ${TARGET_ARCH} instead.
Build-tested and inspected core-image-minimal rootfs with:
* AArch64 no multilib (real loader in correct place)
MACHINE = "qemuarm64"
* AArch64 multilib (symlink in correct place)
MACHINE = "qemuarm64"
MULTILIBS = "multilib:lib32"
DEFAULTTUNE_virtclass-multilib-lib32 = "armv7at-neon"
require conf/multilib.conf
* AArch64 usrmerge (real loader in correct place)
DISTRO_FEATURES += "usrmerge"
MACHINE = "qemuarm64"
* AArch64 multilib usrmerge (symlink in correct place)
DISTRO_FEATURES += "usrmerge"
MACHINE = "qemuarm64"
MULTILIBS = "multilib:lib32"
DEFAULTTUNE_virtclass-multilib-lib32 = "armv7at-neon"
require conf/multilib.conf
* big-endian versions of all of the above by also setting
DEFAULTTUNE = "aarch64_be".
(building glibc only.)
* x86_64 (real loader in /lib as before)[2]
MACHINE = "qemux86"
* x86_64 multilib (real loader in /lib64 as before)
MACHINE="qemux86-64"
MULTILIBS = "multilib:lib32"
DEFAULTTUNE_virtclass-multilib-lib32 = "x86"
require conf/multilib.conf
I also tested leaving an unwanted file in ${exec_prefix}/lib for
do_poststash_install_cleanup to detect, and I believe the detection always
worked correctly.
[1] http://lists.openembedded.org/pipermail/openembedded-core/2018-November/276120.html
(From OE-Core rev: a705c0782c863ee960d65b5109168a4587a0a7b7)
Signed-off-by: Mike Crowe <mac@mcrowe.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
In case the baselib is lib64 we would want to delete /usr/lib
after removing the /usr/lib/locale dir and the implementation
wanted to do that earlier as well but the fault was checking
an already removed dir (/usr/lib/locale) before trying to
remove /usr/lib as that check would always fail.
Now we simply try to delete /usr/lib after deleting
/usr/lib/locale to make sure it deletes cleanly and is empty
at the time of deletion.
(From OE-Core rev: 4dad1568f8f84ec9de4bf7235822f77a8ee6a413)
Signed-off-by: Awais Belal <awais_belal@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
On other distros like ubuntu/centos, libnss-db usually provides:
- The libraries
- The Makefile to create database
(in /var/db for centos, /var/lib/misc/ for ubuntu)
- The makedb command (it's in glibc-common for centos7)
What we had is:
- The libraries are in glibc-extra-nss
- The Makefile is removed
- The makedb command is in glibc-utils (lack of dependency)
So when glibc-extra-nss is installed but glibc-utils is not,
we see error like:
nscd[165]: 165 checking for monitored file `/var/db/group.db': No such file or directory
nscd[165]: 165 checking for monitored file `/var/db/passwd.db': No such file or directory
And there is not an easy way to create these databases.
To fix the issue:
- Re-package the libraries into libnss-db
- Don't remove the Makefile and add it in libnss-db
- Add RDEPENDS for libnss-db on glibc-utils
- Provide a shell script, makedbs.sh, to generate the db files.
This is to avoid dependency on 'make'.
Notes:
1. For external toolchain, an extra package 'libnss-db' need to be provided
If replacing glibc from core.
2. I've check the git history of nss/db-Makefile, the last two functionality
fix is as below.
- fix non-portable `echo -n` usage -- Date: Thu Aug 6 04:14:20 2015 -0400
- Fix db makefile rule for group.db -- Date: Fri Nov 11 14:43:36 2011 +0100
So I think this file is stable enough. And using makedbs.sh which is crafted according
to that file is not likely to cause maintanence problem.
(From OE-Core rev: 13cf502fce8956f95fdc8ac0c7a37d741223bcc9)
Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Drop packaging libcrypt from 2.28+ onwards
We have independent crypt implementation coming from libxcrypt
(From OE-Core rev: 6146b8c4216daf56a69f4e3531861302df6a63a2)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
As reported by ChenQi, leaving bits/wordsize.h out of being
multilibbed introduced a problem in building the SDK for
arm64:
Error: Transaction check error:
file /usr/include/bits/wordsize.h conflicts between attempted installs of lib32-libc6-dev-2.27-r0.armv7vet2hf_vfp and libc6-dev-2.27-r0.aarch64
This effectively reverts commit a74c77d6.
(From OE-Core rev: 90ad502bf8faa233e25cf297c1eeefcb0367aea3)
Signed-off-by: Daniel Díaz <daniel.diaz@linaro.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Once another header #includes <bits/wordsize.h>, there is a
potential recursion going on because the
multilib_header_wrapper.h #includes <bits/wordsize.h> again!
This should not happen because an __arm__ (32-bits) or an
__aarch64__ (64-bits) environment guarantees that we will
be getting the correct definition, but when building against
a different target (like BPF), recursion is what happens.
This can be seen, for instance, when building eBPF programs
from the kernel with `clang -target bpf', such as the ones
located in linux/tools/testing/selftests/bpf/.
(From OE-Core rev: a74c77d6168101e88c3a3bce7130f4f52cfab95d)
Signed-off-by: Daniel Díaz <daniel.diaz@linaro.org>
Signed-off-by: Aníbal Limón <anibal.limon@linaro.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When building SDK via populate_sdk for qemuarm64 with multilib
enabled, we would have conflict about bits/floatn.h at populate_sdk
time.
file /usr/include/bits/floatn.h conflicts between attempted ins
talls of libc6-dev-2.27-r0.aarch64 and lib32-libc6-dev-2.27-r0.armv7vehf_vfp
Apply oe_multilib_header on this header file to fix the problem.
(From OE-Core rev: 650c59c8b6796cf4797ca1860be85f6ccf50bcd2)
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
If ${libdir} is a subdirectory of ${prefix}/lib, e.g. /usr/lib/aarch64-linux, the cleanup logic will delete libc.so.
This bit of code was added in 2012 (git show b744f4cc) to remove /usr/lib/locale, this commit makes it remove that directory recursively and afterwards remove /usr/lib, erroring out if it's non-empty.
Tested with a plain (/usr/lib), a 64-bit (/usr/lib64) and a multiarch (/usr/lib/aarch64-linux) build. I strongly suspect this whole bit of cleanup isn't needed anymore, but my testing is too limited to be certain.
(From OE-Core rev: d8f4c7794f15f7071ee8e621d7964cb4b4134630)
Signed-off-by: Koen Kooi <koen.kooi@linaro.org>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
These functions were moved to meta/lib/oe in 2010 and the base_* functions in
utils.bbclass were intended to be a short-term compatibility layer. They're
still used in a few places, so update the callers to use the new functions.
(From OE-Core rev: c97acbd034532895ce57c6717ed1b3ccc7900b0d)
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Change hardcoded /lib to ${nonarch_base_libdir} to correctly adapt the
code in do_install_append_aarch64() for when usrmerge is enabled in
DISTRO_FEATURES.
(From OE-Core rev: ac373c9f760463d989d6a1eb3a14b7c5b255b9d4)
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We don't have /etc/netgroup by default, so do not
cache for netgroup by default to avoid:
nscd[529]: 529 disabled inotify-based monitoring for file `/etc/netgroup': No such file or directory
nscd[529]: 529 stat failed for file `/etc/netgroup'; will try again later: No such file or directory
(From OE-Core rev: 10007bcd30a96470059f9d5b19cf698243486f06)
(From OE-Core rev: 0adedfc2bf8981819fbbf8b1884da44c7082d1a6)
Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Otherwise it will cause conflicts in mutlilib setting, as it
varies from one machine to another.
(From OE-Core rev: 30140990a8bbe920222204b4f19113f5a1662478)
Signed-off-by: Chunrong Guo <chunrong.guo@nxp.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Add support for armeb of multilib.
(From OE-Core rev: ac3acdcdc313ee13739d378e88b907b31765cfee)
Signed-off-by: zhengrq <zhengrq.fnst@cn.fujitsu.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
uclibc support was removed a while ago and musl works much better. Start to
remove the various overrides and patches related to uclibc which are no longer
needed.
uclibc support in a layer would still be possible. I have strong reasons to
believe nobody is still using uclibc since patches are missing and I doubt
the metadata even parses anymore.
(From OE-Core rev: 653704e9cf325cb494eb23facca19e9f05132ffd)
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>
With this change (combined with the previous linux-libc-header fix), a
combined sysroot for 32 and 64 bit arm works meaning our SDK works
correctly for that multilib setup.
(From OE-Core rev: 4690cd8e34fc23de10400cc1c178b2c73c7690c7)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
If base_bindir=bindir, /usr/sbin/ldconfig gets packaged into glibc-utils instead
of glibc: Switch PACKAGES order to prevent this.
(From OE-Core rev: 2794b1cc12af051db404d46aea4db4c74e8e1fe3)
Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Otherwise it will cause conflicts in mutlilib setting, as it
varies from one machine to another.
(From OE-Core rev: a3b19882e2f50089c785b1e2591550356c12a175)
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
USE_LDCONFIG could previously be set to 0 by distros which do not
require ldconfig or ld.so.conf on the target. Since more and more
recipes may need to respect that option, replace the ad-hoc variable
with a distro feature.
Distros which previously set:
USE_LDCONFIG = "0"
Should now instead use:
DISTRO_FEATURES_BACKFILL_CONSIDERED_append = " ldconfig"
(From OE-Core rev: a905df2dd8f43a2febffa64a39b6e508510326a0)
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>
The change to make do_stash_locale an sstate task between do_install and
do_package has some unforeseen problems since the function deletes part of
${D} but may or may not run depending on whether the task is installed from
sstate.
This cleans up the current "pre packaging" function to be more deterministic
and result in the same set of files, whichever code path we end up reaching
that point by. Its not an ideal sitation but it should avoid the race
failures we've seen on some builds.
(From OE-Core rev: 2a60bba00da9c4e2b1a1ce30e509dba82cc6c94a)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Putting the locale and script files into the sysroot for use by their
specific recipes used to be a simple way to share the files. With RSS,
we don't want to copy these into many different recipes so put these
files in their own sstate task.
(From OE-Core rev: c283f2532851e2a65e08fe7b3fd71c5d9abb70f2)
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>
With the autodebug package generation logic, specifically setting FILES_${PN}-dbg
isn't needed in most cases, we can remove them.
(From OE-Core rev: 3ab59d49dd7c18e194b58d1248b4b87709b5a738)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
$libdir/glibc is deleted if it doesn't exist but this is incorrectly assuming
what variables are used to create this directory. In fact libexecdir is being
used in the Makefile so use that in the recipe too.
This fixes builds where libexecdir is changed.
(From OE-Core rev: 0c73dcb7c1c369a66e4c5804fcdd19f657426a5d)
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
install nscd related configuration file, startup files, and package them,
make nscd easy to startup
(From OE-Core rev: 6d8351ac26295d2e5a693169bd2df95b89cb32fe)
Signed-off-by: Roy Li <rongqing.li@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* INSANE_SKIP_${PN}_append_aarch64 is causing following warning in some
setups:
WARNING: Variable key INSANE_SKIP_${PN} () replaces original key INSANE_SKIP_glibc ().
* in worst case this will be applied also for glibc-initial package
which is using the same glibc-package.inc, but glibc-initial doesn't
create any packages so we should be fine
* someone building for aarch64 should confirm verify that this
INSANE_SKIP is still needed and cannot be fixed properly it was
introduced in:
commit aeb6f53dd607ceb0d2265a05c27f751109c73752
Author: Mark Hatle <mark.hatle@windriver.com>
Date: Thu Dec 18 16:51:13 2014 +0800
glibc-package: aarch64 enable symlink for ABI compliance
aarch64 requires the ld.so to be present in /lib, even if the rest
of the libraries are installed into an alternative directory.
See: https://sourceware.org/glibc/wiki/ABIList
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(From OE-Core rev: 34ffa04a5030d23070aa4d389d1cc51438525670)
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
libmvec is new library in glibc 2.22 and currently turned on by default
on x86_64. this helps in packaging it properly when its generated
Fixes warning like
WARNING: QA Issue: nativesdk-glibc: Files/directories were installed but
not shipped in any package:
/usr/local/oecore-x86_64/sysroots/x86_64-angstromsdk-linux/lib/libmvec-2.21.90.so
Please set FILES such that these items are packaged. Alternatively if
they are unneeded, avoid installing them or delete them within
do_install. [installed-vs-shipped]
(From OE-Core rev: f009e592fd5ef65c2417ec8cc131056a575a003f)
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>
Add functionality to strip binaries/libraries going into the sysroot. Whilst
this does fractionally slow down the build, it also significantly reduces the
size of the sstate cache files.
(From OE-Core rev: 30f3774f4cd5bbb8c1e6884aeff5af91ab053fc1)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
aarch64 requires the ld.so to be present in /lib, even if the rest
of the libraries are installed into an alternative directory.
See: https://sourceware.org/glibc/wiki/ABIList
(From OE-Core rev: aeb6f53dd607ceb0d2265a05c27f751109c73752)
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
- This is a big swoop change where we switch to using glibc
- option-groups are forward ported
- cross-localedef is extracted out from eglibc and hosted
at github.com/kraj/localedef, its used for cross-localedef
recipe
- Other non ported patches from eglibc are forward ported
ppc8xx cache line workaround
SH fpcr values
dynamic resolver
installing PIC archives is there but is not applied
libc header bootstrap
- Delete eglibc recipes we moved back to using glibc now
- Fix ppc/e500 build
- Fix crypt module build when options are used
- Fix fnmatch build when options OPTION_EGLIBC_LOCALE_CODE is unset
HAVE_MBSTATE_T and HAVE_MBSRTOWCS should be defined conditionally
based upon OPTION_EGLIBC_LOCALE_CODE being set/unset
- Move the ports/ patches to relevant files now that ports is gone
(From OE-Core rev: 1027c535ea753e63d9ffe469a423e04467cf8940)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Glibc 2.10.1 is outdated now and eglibc seems to provide a superset of
its functionality.
(From OE-Core rev: 1d9866a2c4fe93202749a56c7ad89a3cbd07d9f8)
Signed-off-by: Phil Blundell <philb@gnu.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
These changes ensure files packaged in the -locale package aren't included
in the main do_install and also ensures the staging directory used for
the -locale package doesn't end up in the -dev package.
(From OE-Core rev: 2dccb370e04bc2bbe208b2b0541077d8ef519535)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
*libc's do_package will cost a lot of time due to the locale handing,
which may delay the other recipe's do_package task and affect the build
performance.
This commit moves locale handling into a separate recipe *libc-locale.
[RP: Add fixup with recent eglibc commit conflict for FILES_pn-dbg and PACKAGES]
(From OE-Core rev: c6d3242f7642f537b02870bb4eb650f0e2094b85)
Signed-off-by: Dongxiao Xu <dongxiao.xu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
[BUGID #281]
Evaluate and update each package in recipes-core to ensure they have a
consistent summary and description.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Along with qemu this provides another option to generate locales for
[e]glibc. The new method is to use cross-localedef with appropriate arch
specific parameters.
The cross-localedef method is found to be 15 times faster than qemu's
emnualted method.
LOCALE_GENERATION_WITH_CROSS-LOCALEDEF : This is new config variable
introduced to selet qemu or cross-localedef method for locale generation.
Thanks to Mark Hatle from Windriver for providing the rich
information for cross locale generation.
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
This commit fixes the following error while doing this:
bitbake poky-image-sato
ERROR: Conflicting PREFERRED_PROVIDER entries were found which resulted in an attempt to select multiple providers (['virtual:nativesdk:/distro/dcui/dexuan/meta/recipes-core/eglibc/eglibc_2.12.bb', '/distro/dcui/dexuan/meta/recipes-core/eglibc/eglibc_2.12.bb']) for runtime dependency glibc-gconv-iso8859-1 The entries resulting in this conflict were: ['PREFERRED_PROVIDER_virtual/libc-nativesdk = eglibc-nativesdk', 'PREFERRED_PROVIDER_virtual/libc = eglibc']
NOTE: multiple providers are available for runtime glibc-gconv-iso8859-1 (eglibc, eglibc-nativesdk, glibc-nativesdk, glibc, external-csl-toolchain, external-poky-toolchain)
NOTE: consider defining a PREFERRED_PROVIDER entry to match glibc-gconv-iso8859-1
And bumped PR.
This partly fixes [BUGID #329]
Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
[e]glibc-nativesdk should not be providing libsegfault.
This commit fixes the following error while doing this:
bitbake poky-image-minimal-initramfs
DEBUG: providers for libsegfault are: ['glibc-nativesdk', 'glibc',
'eglibc-nativesdk', 'external-csl-toolchain', 'eglibc',
'external-poky-toolchain']
DEBUG: selecting
virtual:nativesdk:/rphome/poky/meta/recipes-core/glibc/glibc_2.10.1.bb
as PREFERRED_VERSION 2.10.1 of package glibc-nativesdk (for item
libsegfault)
DEBUG: selecting /rphome/poky/meta/recipes-core/glibc/glibc_2.10.1.bb as
PREFERRED_VERSION 2.10.1 of package glibc (for item libsegfault)
NOTE: checking PREFERRED_PROVIDER_glibc-nativesdk
NOTE: checking PREFERRED_PROVIDER_glibc-nativesdk-2.10.1
NOTE: checking PREFERRED_PROVIDER_glibc-nativesdk-2.10.1-r3
NOTE: checking PREFERRED_PROVIDER_virtual/libc-nativesdk
NOTE: selecting glibc-nativesdk to satisfy runtime libsegfault due to
PREFERRED_PROVIDER_virtual/libc-nativesdk = glibc-nativesdk
And bumped PR.
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
The idea of build-time locale generation is documented in glibc-package.inc:
Binary locales are generated at build time if ENABLE_BINARY_LOCALE_GENERATION
is set. The idea is to avoid running localedef on the target (at first boot)
to decrease initial boot time and avoid localedef being killed by the OOM
killer which used to effectively break i18n on machines with < 128MB RAM.
However it doesn't make sense to do same thing for glibc-nativesdk, as the
build system is powerful. More importantly is that ideally host_arch running
sdk may even be out of the support list of target_arch by qemu-native.
Regarding to above rationale, this commit disables build time locale
generation to avoid following error when asking qemu to run localdef:
NOTE: /opt/poky/sysroots/i586-pokysdk-linux/lib/ld-linux.so.2: No such file or directory
nativesdk binaris have opt path hardcoded to avoid mess with host bits, which
is another reason that build time locale generation is not feasible here.
This fixes [BUGID #264]
also add 'nativesdk' to eglibc per RP's suggestion
Signed-off-by: Kevin Tian <kevin.tian@intel.com>
Other enhancements:
print qemu's stdio & error on failure
glibc: enable locale generation for all arches
eglibc: enable binary locale generation for mips
And cleanup of code based on the code review.
[e]glibc: move common definition in the common file
bitbake was complaining about duplicate definition of
get_libc_fpu_setting in eglibc.inc & glibc.inc
files.
And bump PRs
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
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>
