Where variables are used in python, we need to ensure they are expanded.
This happens to work at the moment but likely will not happen in future
and isn't good code practise.
Its mostly an issue around key values, since bitbake has already
performed key expansion when these functions are executed.
(From OE-Core rev: 6d1142b56948c048111c4f78d9909c1846ab225b)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Configuring gcc with --enable-target-optspace (which causes gcc to
append "-g -Os" to the default CFLAGS_FOR_TARGET and so force libgcc
etc target libraries to always be optimised for size) dates back to
the very first commit in oe-core git in 2005 (for gcc 3.4.3).
Configuring gcc with --enable-target-optspace is not done widely
elsewhere (it's not used for Ubuntu or Fedora host gcc, the Linaro
binary toolchain or in Buildroot since early 2015). Sometime around
gcc 4.5.x it caused problems for powerpc and so was disabled for that
architecture:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=43810
This patch removes --enable-target-optspace completely (ie powerpc is
no longer a special case) and allows optimisation of libgcc etc to be
controlled directly by the flags present in TARGET_CFLAGS.
(From OE-Core rev: 686b266506a1a56fb68ab0f00d658787dd7fe4ce)
Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The Cortex M1, M3 and R4 CPU tuning files are poorly tested (if at
all). They have no obvious users either inside or outside oe-core.
Until OE officially gains support for CPUs without an MMU, these
tuning files are probably better maintained outside of oe-core (e.g.
in a separate meta-nommu layer).
(From OE-Core rev: 7a1445c55de904115b950c8e50432a9f11f02208)
Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
predep_objects and postdep_objects are specifying the duplicate objects
which are better computed by gcc/g++ driver, since we want to generate
PIE, PIC and non-PIC combinations, let gcc decide on this, since it
knows best about linker options to use in each of these situations.
When we defer the linker options to driver, we also need to remove
-nostdlib from linker commandline options.
Remove duplicate directory creation for ${D}${bindir_crossscripts}/
This fixes a frequent problem we see during -fpie links e.g.
|
/home/ubuntu/work/oe/openembedded-core/build/tmp-glibc/sysroots/qemux86-64/usr/include/c++/5.3.0/iostream:74:
undefined reference to `__dso_handle'
|
/home/ubuntu/work/oe/openembedded-core/build/tmp-glibc/sysroots/x86_64-linux/usr/libexec/x86_64-oe-linux/gcc/x86_64-oe-linux/5.3.0/ld:
.libs/cxx_channel.o: relocation R_X86_64_PC32 against undefined hidden
symbol `__dso_handle' can not be used when making a shared object
|
/home/ubuntu/work/oe/openembedded-core/build/tmp-glibc/sysroots/x86_64-linux/usr/libexec/x86_64-oe-linux/gcc/x86_64-oe-linux/5.3.0/ld:
final link failed: Bad value
After this change libtool gets changed as below
old_archive_from_expsyms_cmds=""
# Commands used to build a shared archive.
-archive_cmds="\$CC \$pic_flag -shared -nostdlib \$predep_objects
\$libobjs \$deplibs \$postdep_objects \$compiler_flags \$wl-soname
\$wl\$soname -o \$lib"
-archive_expsym_cmds="\$CC \$pic_flag -shared -nostdlib \$predep_objects
\$libobjs \$deplibs \$postdep_objects \$compiler_flags \$wl-soname
\$wl\$soname \$wl-retain-symbols-file \$wl\$export_symbols -o \$lib"
+archive_cmds="\$CC \$pic_flag -shared \$predep_objects \$libobjs
\$deplibs \$postdep_objects \$compiler_flags \$wl-soname \$wl\$soname -o
\$lib"
+archive_expsym_cmds="\$CC \$pic_flag -shared \$predep_objects
\$libobjs \$deplibs \$postdep_objects \$compiler_flags \$wl-soname
\$wl\$soname \$wl-retain-symbols-file \$wl\$export_symbols -o \$lib"
# Commands used to build a loadable module if different from building
# a shared archive.
@@ -11908,8 +11908,8 @@
# Dependencies to place before and after the objects being linked to
# create a shared library.
-predep_objects="/home/ubuntu/work/oe/openembedded-core/build/tmp-glibc/sysroots/qemux86-64/usr/lib/../lib/Scrt1.o
/home/ubuntu/work/oe/openembedded-core/build/tmp-glibc/sysroots/qemux86-64/usr/lib/../lib/crti.o
/home/ubuntu/work/oe/openembedded-core/build/tmp-glibc/sysroots/qemux86-64/usr/lib/../lib/x86_64-oe-linux/5.3.0/crtbeginS.o"
-postdep_objects="/home/ubuntu/work/oe/openembedded-core/build/tmp-glibc/sysroots/qemux86-64/usr/lib/../lib/x86_64-oe-linux/5.3.0/crtendS.o
/home/ubuntu/work/oe/openembedded-core/build/tmp-glibc/sysroots/qemux86-64/usr/lib/../lib/crtn.o"
+predep_objects=""
+postdep_objects=""
predeps=""
postdeps="-lstdc++ -lm -lgcc_s -lgcc -lc -lgcc_s -lgcc"
(From OE-Core rev: e1b45cb8e15458dde21f5e86ea1cbb619fb5c4cf)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Create a link to stylesheets directory for docbook-xsl-stylesheets to
make it easy to refer.
(From OE-Core rev: 15531f8ba83174bc36aa2566721dd80188a8d465)
Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When using -fpie (security related cflags), qemux6-64 builds would
fail due to linker symbol errors. This was due to the internal zlib
that binutils was using.
Add the switch to tell it to use the target system zlib which is
already in DEPENDS. That zlib is already compiled with the correct
flags.
(From OE-Core rev: 9ec713c3949f9bcfa1e03561b9811eb3d0451c2f)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
G4 does not have SPE, so we make that explicit in the tune files and
since we emulate G4 when building Qemu, we ensure it for qemuppc as
well.
GCC config for powerpc-linux is made to include SPE by default which is
equivalent if the tripet was powerpc-linux*spe, this forces gcc to
configure assembler to enable -mspe by default, when we do that then the
kernel fails to compile with binutils 2.26, since newer assembler is
smart to detect the tlbia instructions are not compatible with SPE and
hence the kernel build breaks rightly. We configure the kernel for G4 as
well where it enables tlbia instrucitons rightly so because it thinks
its being configured for power4. So we keep the options but do not force
-mspe down to assembler as default.
(From OE-Core rev: 7a51776a830167e43cbd185505f62f328704e271)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When compile with readline 5.2:
completer.o: In function `gdb_display_match_list':
completer.c:(.text+0x1c13): undefined reference to `_rl_completion_prefix_display_length'
completer.c:(.text+0x1ce8): undefined reference to `rl_sort_completion_matches'
collect2: ld returned 1 exit status
The --without-system-readline will make it work.
(From OE-Core rev: 94e1b917078bedf73830e54278af77f742c93581)
Signed-off-by: Robert Yang <liezhi.yang@windriver.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>
Remove mempcpy patch its already in uclibc now
refresh libargp patch to match latest elfutils
Add ALLOW_EMPTY for musl
core-image-cmdline-full asks for elfutils package to be added to package
it has all eu-* tools but we dont build them on musl. nevertheless lets
satisfy the dependency so image can build for musl
Build libdw,libasm,libebl along with libelf
enhance the musl portability patch to compile every other library
except the eu-* tools
Fix build on aarch64/musl issues
Header inclusion ends up with duplicate definitions of iovec struct
since uclibc defines the same in usersapce, we have been depending upong
including the kernel header on aarch64, using sys/uio.h is the right
approach
Fix x86 build warning seen on musl
(From OE-Core rev: a28c70455931e0b7e3449a76c3b9132bd572478b)
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>
mode_t is needed by gdbserver headers so include sys/types.h to get them
(From OE-Core rev: b0742a41e80c48de98f82f7d1e963d462c941934)
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>
Both e2fsprogs and util-linux can build blkid, but we want to always use
util-linux's for consistency.
(blkid was enabled again in a1f235)
(From OE-Core rev: eaff837048bc4fba0ed64fd612634682fe254d7f)
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The cross-canadian compiler needs the nativesdk compiler to build
but for some reason this was missing. Add the missing dependency.
(From OE-Core rev: 77fcdfdc83f01f1ff0ae310ebb7bffbdb4330156)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This reverts commit 5c69c561a76cb10d7896ae0a0399190f11b2e0ca.
The change was incomplete, not handling cases such as the fetcher
using xz, or linux-yocto, dpkg or apt or the lzma image type.
(From OE-Core rev: fc4209baa098caebf9c4cb75f9a6f2e85f43333c)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The old entropywave.com URL no longer responds.
Also remove unnecessary 'name=orc' from SRC_URI.
(From OE-Core rev: 79b519e8622e9bffc79299dee096a85fd4ac2909)
Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
At some point we may want to add xz-native to ASSUME_PROVIDED. This
allows that to work whilst still allowing access to liblzma for
those things which need it (e.g. pixz).
(From OE-Core rev: 5c69c561a76cb10d7896ae0a0399190f11b2e0ca)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
strol_l is not available on musl
delete charset.alias as well its not needed
Use internal gc function on musl
We get errors on certain functions not being available in boeheme gc
when built with musl. Therefore use the internal versions
e.g.
Undefined gc_set_finalizer_notifier
(From OE-Core rev: 31ee3460c3a0578ed36f3f5a940617ccb873c721)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
enable largefile support if its in DISTRO_FEATURES
(From OE-Core rev: 300defd71205bf350a6e956451e0cc87c4e92902)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Turn bzip2 support into a PACKAGECONFIG
zlibs is a must have and therefore add it to DEPENDS will make the build
consistent
(From OE-Core rev: ded03dab901647a7388d219a3a228d4f4f4298e8)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Remove a glibc specific patch to be applied only for native version,
this is a dpkg-native specific patch needed for Centos 5.8, this patch
should be redone to be glibc independent for it to be applied to all
types of dpkg recipes
(From OE-Core rev: 610a9f102a83e1ffb9d7954e40df5f65d24d3a26)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
unlike glibc, uclibc does not have internal APIs implemented for
libiconv, so it needs to depend upon libiconv
(From OE-Core rev: 04a609a70f6c464ffd32a4a073c3801d9446622b)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Replaced by 0002-remove-rpath.patch during the 3.10 -> 3.11 update.
(From OE-Core rev: aab27d556adf15e05351ec1a9a3c8ccc51717c75)
Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The sh-mem-random.c test app tries to use neon loads and stores to
test 64-bit float copies when building for ARM. Allow it to do so if
possible, but fallback to C when building for ARM targets which don't
support neon.
(From OE-Core rev: 2194bbfe287412449d85182a4e02c8884c83170f)
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 intdiv test has been partially fixed upstream and the
vcvt_fixed_float_VFP test can be fixed with a similar approach, ie
ensuring that it is always compiled with appropriate
-march/-mcpu/-mfpu flags to support the instructions being tested.
For tests requiring armv7ve instructions, ensure that we set both
-march=armv7ve and -mcpu=cortex-a15 (since some TUNE_CCARGS may set
-march=armv7-a and adding -mcpu=cortex-a15 alone is not enough to
over-ride that).
See similar cases in none/tests/arm/Makefile.am
(From OE-Core rev: 3e94af90d90f6db6bab66ae87a47b31f3a0474f6)
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 python-nose upgrade to 1.3.7 required a new unittest runtime dependency
after this, bitbake cant find a runtime provider for it when trying to build
python-nose-native, this patch makes python-native RPROVIDES python-unittest-native
to fix the issue.
(From OE-Core rev: 4583cd1bb15306e8f0ab7bcd80732e6f35aa4533)
Signed-off-by: Alejandro Hernandez <alejandro.hernandez@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Noteworthy changes in release 4.11 (2015-12-21)
===============================================
* Changes in behavior
* Some syscalls have different names depending on architecture and
personality. In these cases, syscall filtering and printing now
consistently uses the names that match corresponding __NR_* kernel
macros of the tracee's architecture and personality.
* Added strace prefix to all diagnostic messages.
* Improvements
* Enhanced and extended test suite.
* Implemented reliable tracing of processes whose personality differs
from the personality of strace, fixed decoding of 32-bit personality
syscalls on 64-bit architectures.
* When some data referenced by syscall arguments is irrelevant or cannot
be read from tracee's memory, strace now prints its address consistently.
* When a syscall is restarted using restart_syscall mechanism, strace now
shows the syscall name on architectures that expose this information.
* Print values returned by epoll_create1, eventfd, eventfd2, fanotify_init,
inotify_init1, perf_event_open, pipe, pipe2, signalfd, signalfd4, socket,
socketpair, timerfd, and timerfd_create syscalls, as well as F_DUPFD and
F_DUPFD_CLOEXEC fcntl commands, as file descriptors.
* Optimized decoding of indirect socket syscalls.
* Implemented decoding of nanoseconds along with seconds in stat family
syscalls (addresses Fedora bug #1251176).
* Implemented decoding of struct statfs.f_flags in statfs syscall.
* Implemented decoding of siginfo_t.si_syscall and siginfo_t.si_arch
in SIGSYS signal.
* Implemented decoding of indirect syscall on mips o32.
* Implemented decoding of IPPROTO_IP control messages.
* Implemented decoding of setsockopt syscall multicast arguments.
* Enhanced decoding of struct sigevent.
* Enhanced decoding of block, loop, mtd, ptp, rtc, SG_IO, socket, tty,
and v4l2 ioctl syscall arguments.
* Enhanced decoding of epoll_create, fcntl, fcntl64, futex, getdents,
getdents64, getsockopt, io_*, mount, msgctl, prctl, quotactl, recvfrom,
sendfile, setsockopt, and shmctl syscalls.
* Implemented decoding of bpf, execveat, ioperm, iopl, kcmp, kexec_file_load,
lookup_dcookie, membarrier, memfd_create, mlock2, name_to_handle_at,
open_by_handle_at, sched_getattr, sched_setattr, and userfaultfd syscalls.
* Updated lists of AF_*, EPOLL_*, FALLOC_FL_*, F_*, INPUT_PROP_*, IPPROTO_*,
IPV6_*, IP_*, KEY_*, LOCK_*, MNT_*, MS_*, PACKET_*, PERF_FLAG_*, PF_*,
PR_*, PTRACE_O_*, RENAME_*, SCTP_*, SECBIT_*, SO_*, TCP_*, and siginfo_t
constants.
* Added Nios II architecture support.
* Added new syscall entries to match Linux 4.4.
* Bug fixes
* Fixed potential short read of strings from tracee's memory.
* Fixed -qq option in conjunction with -o option.
* Fixed filtering of <unavailable> syscalls.
* Fixed decoding of ioctl syscall command on aarch64 and 32-bit architectures
with 64-bit aligned structures.
* Fixed decoding of eventfd2, epoll_create1, signalfd4, and sync_file_range2
syscall flags arguments.
* Fixed decoding of 4th argument of clock_nanosleep syscall.
* Fixed decoding of getpagesize syscall on m68k.
* Fixed decoding of getrandom and seccomp syscalls on aarch64.
* Fixed decoding of timezone argument of gettimeofday and settimeofday
* syscalls.
* Fixed decoding of offset argument of mmap2 syscall on arm, sparc,
and sparc64.
* Fixed decoding of pipe syscall on alpha and mips.
* Fixed decoding of pipe2 syscall on ia64, sh, sparc, and sparc64.
* Fixed decoding of signal mask argument of pselect6 syscall on x32.
* Fixed decoding of signal mask of rt_sigreturn syscall on aarch64, m68k,
and x32.
* Fixed decoding of 4th argument of semctl indirect syscall.
* Fixed decoding of sa_restorer and sa_mask of sigaction syscall on m68k.
* Fixed decoding of statfs64 and fstatfs64 syscalls on arm eabi.
* Fixed decoding of struct dirent on x32.
* Fixed decoding of times syscall return value on mips n32 and x32.
* Fixed decoding of path argument of umount2 syscall
(addresses Debian bug #785050).
* Worked around a kernel bug in tracing privileged executables.
* Fixed various errors in mapping between syscall numbers and associated
information like the number of syscall arguments, the name of syscall,
and syscall decoder.
(From OE-Core rev: f3e6bf135b7b69614d1405ad99e1264ea70a9197)
Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
TARGET_ARCH doesn't map to strace's ARCH variable, so we can't assign
one from the other. Forcing the incorrect value via the make command
line doesn't cause any problems with strace v4.10, but it will do for
strace v4.11.
(From OE-Core rev: b4bdbdd65dd9768209f453f1c2eacf1f7f78a998)
Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
