ARM stack frames for abort and raise were limited to the the actual
abort and raise call, such as:
Obtained 4 stack frames.
./test-app(print_trace+0x1c) [0x10a08]
./test-app() [0x10b3c]
/lib/libc.so.6(__default_sa_restorer+0) [0x4adae1e0]
/lib/libc.so.6(gsignal+0xa0) [0x4adacf74]
This is not terribly useful when trying to figure out what function
may have called called the abort, especially when using pthreads.
After the change the trace would now look like:
Obtained 8 stack frames.
./test-app(print_trace+0x1c) [0x10a08]
./test-app() [0x10b3c]
/lib/libc.so.6(__default_sa_restorer+0) [0x4befe1e0]
/lib/libc.so.6(gsignal+0xa0) [0x4befcf74]
/lib/libc.so.6(abort+0x134) [0x4befe358]
./test-app(dummy_function+0x50) [0x10adc]
./test-app(main+0xd4) [0x10c24]
/lib/libc.so.6(__libc_start_main+0x114) [0x4bee7a58]
(From OE-Core rev: 93bf8713d8e13c278543baea94fb8dad0cb80e49)
Signed-off-by: Yuanjie Huang <yuanjie.huang@windriver.com>
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Backport patch to fix CVE-2016-6323 of glibc. And remove the section of
file ChangeLog which can't be applied.
Ref:
https://sourceware.org/bugzilla/show_bug.cgi?id=20435
(From OE-Core rev: a3c2acee40c8875e311e03bff6906e7c93c491fc)
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>
Reduce duplication in MIPS variants now that the MACHINEOVERRIDES
variable is defined
(From OE-Core rev: 5013504b8b406b47a66c84413cb7f8beb5c336c3)
Signed-off-by: Zubair Lutfullah Kakakhel <Zubair.Kakakhel@imgtec.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Add support for MIPS Release 6 ISA
The loader is located at a new place for multiarch.
For more details, check https://wiki.debian.org/Multiarch
and https://sourceware.org/glibc/wiki/ABIList#mips
(From OE-Core rev: 502145553e2368525a4b327c64d2971dac58d85b)
Signed-off-by: Zubair Lutfullah Kakakhel <Zubair.Kakakhel@imgtec.com>
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>
The ELF specification indicates symbol resolution should be breadth first, not
depth first.
The dl-deps.c: dl_build_locale_scope function is processing in a depth first
mode. This is causes certain symbols to be incorrectly reported when
LD_TRACE_PRELINKING=1 is enabled.
See glibc BZ #20488 for more information.
(From OE-Core rev: fb72263eaa94e64ddeee457b5b1bc999f0e647da)
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
glibc 2.24 is released now
https://www.sourceware.org/ml/libc-alpha/2016-08/msg00212.html
(From OE-Core rev: 01fe48252085284e2964f5dd52b8b5fa54ee10d3)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The nativesdk libc when used by buildtools has a hard requirement on supporting
a UTF-8 locale because Python 3 needs a UTF-8 locale. However we currently only
ship the C locale, which means that Python attempts to lookup the user's locale
(for example, en_NZ.UTF-8) in the locale archive under it's prefix it fails and
falls back to C. This the results in Python using ASCII instead of UTF-8 for
file encoding, and bitbake breaks.
Th obvious solution would be to ship all locales, but this would add
approximately 250MB to the size of the buildtools tarball (which is currently
around 30MB). Generating a binary locale archive reduces this down to 100MB,
but this is still a drastic increase in footprint. If we ship a subset of
locales in the tarball then there will be users whose locale isn't in the
tarball, and they'll have to change their locale to an "approved" one, which
isn't the best of messages to send to new users.
The alternative is to tell the nativesdk libc that the locale archive isn't
under it own prefix but is in fact at /usr/lib/locale/locale-archive, so the
buildtools libc uses the host locale archive. The locale archive format appears
to be at least fairly stable: our glibc 2.24 can read the locale archive
generated by glibc 2.17 (Centos 7).
[ YOCTO #9775 ]
(From OE-Core rev: 75321b6b0f2c0ac667b9350b387b01a188e195c8)
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
glibc master added the EM_METAG tag but didn't add the relocation defines.
However the kernel tooling only checks for EM_METAG when defining its own values
so scripts/recordmcount ends up using R_META_* symbols without their definition.
Whilst the kernel can and should be fixed, this breaks all users of recordmcount
so patch elf.h to add the values.
(From OE-Core rev: 61f73ae289bf8dfe72d5f4beaac966fb4ac8dc90)
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
- libc-package.bbclass: Do not use --old-style
This option has been dropped from latest glibc
(From OE-Core rev: 78ab1e7cdedc6a73395af5d053b49cf081416732)
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>
Currently locales are not found in a relocated buildtools-tarball
such as that used in eSDK. This breaks bitbake when used under python3.
This patch adds enough relocation magic to nativesdk-glibc so that
the binary locales can be found even in a relocated buildtools-tarball
and bitbake works successfully under python3. The eSDK also works
correctly after this change too.
(From OE-Core rev: faa1229ba848e7a4a90638a98d6c809065016a93)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
python3 has much stricter locale requirements than previous versions.
If a locale isn't present, python3 reverts to ascii mode under which
bit
bake can't operate.
We therefore need working binary locales in things like uninative-tarball and buildtools-tarball. This patch enables binary locales for nativesdk-glibc.
(From OE-Core rev: 6993d18062bcc20713dd39399bd7cb3623853c75)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This hunk of patch went missing during the glibc upgrade to 2.24
and without it, uninative-tarball doesn't work properly. This adds
it back so we can spin a new version of uninative.
(From OE-Core rev: a76cbbd5633301dfcd29f18545e1e6f32e3b2448)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Update so this works with python3.
(From OE-Core rev: 20190566db6d77ee0ccd799587db3dfa35e8029a)
Signed-off-by: Tim Orling <timothy.t.orling@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Use the autotools default configure commands and just tell autotools
where to run configure from.
This fixes the build when running on an aarch64 host, which the prebuilt
configure scripts with glibc 2.22 do not recognize.
(From OE-Core rev: 33d4c758a5d71435437dde74556d32404d91342f)
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
They were CEVs and should be already in the source after upgraded.
(From OE-Core rev: e8a5332d467434ee65e0f29927abb9c51b025aff)
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Yocto's ldconfig-native was exported from an old version of glibc and
generates an ld.so.cache that is partially ignored by current versions
of glibc when loading libraries. This occurs for 64-bit binaries, where
glibc expects FLAG_ELF_LIBC6, but ldconfig-native only generates the
standard ELF flag. Fix this with an additional patch on top of the patch
for now.
You can verify this by applying the patch below to your target copy of
glibc and running
LD_DEBUG=libs /lib64/ld-linux-x86-64.so.2 --list $anybinary
--- ./elf/dl-cache.c.orig 2016-01-07 11:05:36.823447171 +0100
+++ ./elf/dl-cache.c 2016-01-07 11:19:53.925878628 +0100
@@ -106,6 +106,8 @@
if (_dl_cache_check_flags (flags) \
&& _dl_cache_verify_ptr (lib->value)) \
{ \
+ if (__glibc_unlikely (GLRO_dl_debug_mask & DL_DEBUG_LIBS))\
+ _dl_debug_printf (" considering file %s\n", cache_data + lib->value); \
if (best == NULL || flags == GLRO(dl_correct_cache_id)) \
{ \
HWCAP_CHECK; \
@@ -117,6 +119,9 @@
searching. */ \
break; \
} \
+ } else { \
+ if (__glibc_unlikely (GLRO_dl_debug_mask & DL_DEBUG_LIBS)) \
+ _dl_debug_printf (" ignoring file %s due to flags %x, expected %x\n", cache_data + lib->value, lib->flags, GLRO(dl_correct_cache_id)); \
} \
} \
while (++middle <= right); \
@@ -265,14 +270,23 @@
/* Only accept hwcap if it's for the right platform. */
#define HWCAP_CHECK \
- if (lib->hwcap & hwcap_exclude) \
+ if (lib->hwcap & hwcap_exclude) { \
+ if (__glibc_unlikely (GLRO_dl_debug_mask & DL_DEBUG_LIBS)) \
+ _dl_debug_printf (" hwcap mismatch %x vs. %x\n", lib->hwcap, hwcap_exclude); \
continue; \
- if (GLRO(dl_osversion) && lib->osversion > GLRO(dl_osversion)) \
+ } \
+ if (GLRO(dl_osversion) && lib->osversion > GLRO(dl_osversion)) { \
+ if (__glibc_unlikely (GLRO_dl_debug_mask & DL_DEBUG_LIBS)) \
+ _dl_debug_printf (" os version mismatch %x vs. %x\n", lib->osversion, GLRO(dl_osversion)); \
continue; \
+ } \
if (_DL_PLATFORMS_COUNT \
&& (lib->hwcap & _DL_HWCAP_PLATFORM) != 0 \
- && (lib->hwcap & _DL_HWCAP_PLATFORM) != platform) \
- continue
+ && (lib->hwcap & _DL_HWCAP_PLATFORM) != platform) { \
+ if (__glibc_unlikely (GLRO_dl_debug_mask & DL_DEBUG_LIBS)) \
+ _dl_debug_printf (" platform mismatch %x vs. %x\n", lib->hwcap & _DL_HWCAP_PLATFORM, platform); \
+ continue; \
+ }
SEARCH_CACHE (cache_new);
}
else
This version of ldconfig-native should really be replaced with a version
matching the glibc source in use on the target platform.
(From OE-Core rev: bf9c1e6fa1c8eb86670383bad9b7c2e54bfe17f4)
Signed-off-by: Clemens Lang <clemens.lang@bmw-carit.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We're seeing various failures where nativesdk glibc can't find the gconv
modules. We've tried various workaround but this needs fixing 'properly'.
One significant problem is knowing when a binary would use this glibc
and hence when to set the path and when not to.
Add the default path to its own special section which the sdk
relocation script already knows how to handle and remap.
This significantly improves the usability of uninative.
(From OE-Core rev: d40cb4a359dbc5d762fabd996c29e468f5398dd9)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
- Avoid code duplication to handle -O, -O1 and -Os cases
- Consider the effective optimisation level only (avoids spurious
warnings if multiple optimisation flags are present).
- Prefix warnings with PN instead of hardcoding "glibc" (avoids
confusing warnings since the test is also applied to glibc-initial,
nativesdk-glibc, nativesdk-glibc-initial, etc, and each could
potentually have different optimisation flags).
(From OE-Core rev: 9d72015e3458c78a9f0d20ad3dc27c8a9bb1069c)
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>
Drop kconfig and options-group support
Forward port cross-localedef support
Assume ssp support in libc when building gcc-initial
(From OE-Core rev: 9c3d461c4d54d684b38ec4c038a1c3c2fb9923f0)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
it was found that out-of-range time values passed to the strftime function may
cause it to crash, leading to a denial of service, or potentially disclosure
information.
(From OE-Core rev: cf747f0bbcd53af41a7f3981ac65c2b6b6e668f8)
Signed-off-by: Armin Kuster <akuster@mvista.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
A stack overflow vulnerability was found in nan* functions that could cause
applications which process long strings with the nan function to crash or,
potentially, execute arbitrary code.
(From OE-Core rev: 2859c809ac59789d0e631b5386bdf78a5d5c4f4c)
Signed-off-by: Armin Kuster <akuster@mvista.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
A stack overflow vulnerability in the catopen function was found, causing
applications which pass long strings to the catopen function to crash or,
potentially execute arbitrary code.
(From OE-Core rev: e0f71f123147bf4f48cc90c7f26a50164ed4115e)
Signed-off-by: Armin Kuster <akuster@mvista.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The process_envvars function in elf/rtld.c in the GNU C Library (aka glibc or
libc6) before 2.23 allows local users to bypass a pointer-guarding protection
mechanism via a zero value of the LD_POINTER_GUARD environment variable.
(From OE-Core rev: bc51411d2edda908cbef733066d78a986dfec0c0)
Signed-off-by: Armin Kuster <akuster@mvista.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
WARNING: QA Issue: glibc-locale: /glibc-binary-localedata-sd-in/usr/lib/locale/sd_IN/LC_CTYPE is owned by uid 1000, which is the same as the user running bitbake. This may be due to host contamination [host-user-contaminated]
fix type
(From OE-Core rev: d53a0d410ae2390528f1090a131cd3e16d7ebcc5)
Signed-off-by: Armin Kuster <akuster@mvista.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Add patch from commit 96b1b5c127e9e0e637aaf7948cf3330a94a5cd57 to cross-localedef-native
to avoid broken images built with ENABLE_BINARY_LOCALE_GENERATION set to 1:
$ sh -c "export LANG=de_DE; ls -la"
sh: loadlocale.c:130: _nl_intern_locale_data: Assertion `cnt < (sizeof (_nl_value_type_LC_COLLATE) / sizeof (_nl_value_type_LC_COLLATE[0]))' failed.
Aborted
(From OE-Core rev: be456fb6160bf4ef0e90ebe2014a349318d58310)
Signed-off-by: Jens Rehsack <sno@netbsd.org>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We need to expand SELECTED_OPTIMIZATION later, so
do not immediate expansion, and do the work in
anonymous python function.
It is reasonable to give an error for -O0.
[YOCTO #7058]
(From OE-Core rev: 6fef0d528158e6317793b1ebfa26a7b5413c3e80)
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.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 glibc do_compile() environment.
(From OE-Core rev: 5865fdc103fb92007f00a238b925be43af1378f5)
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>
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>
It turns out writing the same list of packaging tasks multiple
times in multiple places is error prone. Move this to a new class
'nopackages", migrate existing users and add glibc-initial and
libgcc-initial since we don't want packages for those recipes.
This means the sstate for those recipes won't be installed, saving
small amounts of build time and bandwidth.
A reference to the old package_write task is also dropped.
(From OE-Core rev: cece583d58f82a50c3a4eac876eb326ac3b8f5e5)
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>
When confronted with an empty line, getmntent() can underrun
a buffer, possibly doing very strange things if it finds
additional space/tab characters. Backport the upstream fix.
(From OE-Core rev: 983a19a65a31b54a6f505181012bd311c28a0ae1)
Signed-off-by: Peter Seebach <peter.seebach@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
A bug in glibc 2.22's ld.so interface for the prelink support causes
the displayed values to be incorrect. The included path fixes this
issue.
Clear ELF_RTYPE_CLASS_EXTERN_PROTECTED_DATA for prelink
prelink runs ld.so with the environment variable LD_TRACE_PRELINKING
set to dump the relocation type class from _dl_debug_bindings. prelink
has the following relocation type classes:
where ELF_RTYPE_CLASS_EXTERN_PROTECTED_DATA has a conflict with
RTYPE_CLASS_TLS.
Since prelink doesn't use ELF_RTYPE_CLASS_EXTERN_PROTECTED_DATA, we
should clear the ELF_RTYPE_CLASS_EXTERN_PROTECTED_DATA bit when the
DL_DEBUG_PRELINK bit is set.
(From OE-Core rev: 12c86bdcc60c54e587a896b0dceb8bb6cc9ff7e3)
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The fix consist of allowing 64bit atomic ops for x86.
This should be safe for i586 and newer CPUs.
It also makes the synchronization more efficient.
[YOCTO #8140]
(From OE-Core rev: 2b8c7aa51f6ac7f79c4834e04b697c04afc8beaf)
Signed-off-by: Juro Bystricky <juro.bystricky@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The 'BRANCH' variable name has no explicit relation with the
SRC_URI. Using 'SRCBRANCH' makes it more obvious and easier to
identify.
This patch makes the use consistent across the metadata.
(From OE-Core rev: ed86bf9c327ceda3976e799ca453028382f277a7)
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Switching MACHINE was causing nativesdk-glibc to rebuild. This was
from the use of OVERRIDES in one of the functions. Exclude OVERRIDES
from the checksum to avoid this.
[patch to oe-selftest to ensure this doesn't regress follows]
(From OE-Core rev: 5e8993632e48e83aa4bab90363506b9d5f7a468f)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The nscd init script uses #! /bin/bash but only really uses one bashism
(translated strings), so remove them and switch the shell to #! /bin/sh.
(From OE-Core rev: 487d4b4d5521ca490e22668ca66921504a1b898f)
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Fix a variety of problems such as typos, bad punctuations, or incorrect
Upstream-Status values.
(From OE-Core rev: bd220fe6ce8c3a0805f13a14706d3130ea872604)
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
provide the nfs-client IMAGE_FEATURES, to ease a user to only
install nfs client related files to image
(From OE-Core rev: bbdc808a8ea76369a36473c906991e25ca49c323)
Signed-off-by: Roy Li <rongqing.li@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Yuanjie Huang
Kai Kang
Zubair Lutfullah Kakakhel
Juro Bystricky
Markus Lehtonen
Jérémy Rosen
Jagadeesh Krishnanjanappa
Mark Hatle
Khem Raj
Ross Burton
Richard Purdie
Tim Orling
Tristan Van Berkom
Robert Yang
Clemens Lang
Andre McCurdy
Armin Kuster
Jens Rehsack
Hongxu Jia
Alexander Kanavin
Peter Seebach
Li Xin
Otavio Salvador
Roy Li