With --force-overwrite (implied by --force-all), dpkg will not abort
when a package overwrites files from different packages. As this can
also lead to "The following package disappeared from your system as
all files have been overwritten by other packages: <package>" and
subsequently broken dependencies, this makes the simple case of
conflicting files hard to debug.
Instead of finding all possibly required force options, only disable
overwrite for now.
(From OE-Core rev: b57b311cfaabbcd08ce51760abcd64cf23e1435b)
Signed-off-by: Jan Luebbe <jlu@pengutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The specific issue here is rpc/rpc.h, but its likely more general.
/usr/include is searched for rpc/rpc.h and if it exists on the
system, it changes behavior. If you are using the extended buildtools
tarball on a machine that has /usr/include/rpc/rpc.h, it will decide
that is good enough and not continue to search. nis fails to build
because /usr/include and /usr/lib are not part of the include/link
paths for the buildtools tarball compiler(nor should they be).
This makes it so python3-native will not build if you are using the
extended buildtools tarball, but from a larger issue perspective it
is building in likely different ways depending on what machine it
is building on.
libtirpc is already a depend so we shouldn't need the hosts rpc/rcp.h.
(From OE-Core rev: f37dfc7907ae7bac08d40468ddde2e5b8bba030c)
(From OE-Core rev: db8bef336849570dd99900b173c44cc47b208058)
Signed-off-by: Jeremy A. Puhlman <jpuhlman@mvista.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 400743867d)
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
pseudo-native is special in that bitbake ends up executing it from the
sysroot-components directory before we have any workdirs for the bitbake
fakeroot worker. Since we switched to dynamically linking sqlite, it
means sqlite from the host system may be found, we really want the version
in sysroot-components. Trying to run tasks to create some special environment
for pseudo is hard and error prone. The simplest fix is to add an RPATH to
the binary so that it can correctly find the sqlite we want.
Unfortunately passing $ORIGIN into make doesn't work so well with shell
quoting so we have to fix that during do_install.
[YOCTO #13814]
(From OE-Core rev: 2cbb3a663a3d3b2dcf8c12c26843d6087789ded0)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 3937ca9e2dfabb1ce9bce1d536b60b1e2a43739b)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
THE LICENSE checksum changed in this update due to copyright notice
added for 2020.
(From OE-Core rev: 3c40cfe7433999272e1698e2c914d6d190f76b63)
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Prevent the full recipe-sysroot path from leaking into configargs.h. The
configargs.h header is intended to be static and unchanged as the
content is used as a means of determining that a gcc plugin is built for
the same gcc. This also effects the output of 'gcc -v'. Due to per
recipe sysroots and staging, the sysroot path would be replaced with the
sysroot local to the recipe thus changing the content of configargs.h.
This change also improves gcc binary reproducibility. The sysroot path
is replaced with the base target root "/".
(From OE-Core rev: 0f418fccab3f67a2afaa28195263d6f24831dd56)
Signed-off-by: Nathan Rossi <nathan@nathanrossi.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit b8d6e2ab68ee5e341fe970b191bfd334e6d2c40b)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Prevent the native(sdk) sysroot path from leaking into configargs.h. The
configargs.h header is intended to be static and unchanged as the
content is used as a means of determining that a gcc plugin is built for
the same gcc. This also effects the output of 'gcc --version'. Due to
per recipe sysroots and staging, the sysroot path would be replaced with
the sysroot local to the recipe thus changing the content of
configargs.h.
The sysroot path is replaced with a generic "/host" prefix which
represents the host sysroot (e.g. native or nativesdk).
(From OE-Core rev: 9bb270b3f12ff94b1541649078741e683020ffe9)
Signed-off-by: Nathan Rossi <nathan@nathanrossi.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
(cherry picked from commit 84a78f46d59447eeec3d69532a7506148f64c979)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When configuring the cross-candian toolchain for a non-linux target system,
the resulting gcc configuration included:
--enable-initfini-array--without-headers
these should have been two separate options.
(From OE-Core rev: fdd3d65b690c9d460a5758cf1b83e7b2edfc9559)
Signed-off-by: Mark Hatle <mark.hatle@kernel.crashing.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 7b52893632dae7bc9ac75dddc7ad625e19f41050)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Add a missing makefile dependency which can cause differences in
configuration (submitted upstream).
[YOCTO #13800]
(From OE-Core rev: 1589115ff42e8c211bc0784bd8aca7d3a5b8f566)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit fe97845a45434902c5a994e253a127a462d7d3b4)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The code is encoding host compiler parameters into target builds. Avoid
this for our target builds. This should resolve builds which aren't
reproducible between hosts with different compilers.
(From OE-Core rev: 4b41afeea632f33a490d75621e2c0d6bb2bb6aca)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 71cdbf426e46e3ca1b5038f40e9f7ba958abc537)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Occasionally the reproducibility selftest fails because perl-pod differs,
specifically that the perltoc.pod file was sometimes missing modules.
Debugging revealed that there are missing dependencies so there is a build race:
building perltoc.pod from an clean build tree results in no modules being listed
at all.
A bug has been filed at https://github.com/arsv/perl-cross/issues/86 to solve
this properly, but for now we can just delete perltoc.pod after make has
finished and re-generate it.
[ YOCTO #13726 ]
(From OE-Core rev: 4ee9c60797e95674ae138245b3a4de063b2e95db)
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 7e0f6c9ea4f824f29dc30c6631fd8039ebe83a0b)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Particularly remove a broken detection of skipped tests
(which was marking tests that actually passed), and install
additional artefacts needed for testing.
(From OE-Core rev: 559fd361440898479937c887a0e4f8cfb8c46891)
Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit dbbce0918617c21d0e43e9727d38372c22dff3dc)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Modules actually make use of these files, so they belong in perl-core
other than perl-doc (the immediate failure was ptests for
libmodule-build-perl failing).
(From OE-Core rev: c95832dbfe4f375dadbffc8ee678052dbf7b3913)
Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 829e8c49833e4cb8de5db869769eb492c827e8c9)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Otherwise, some modules such as libmodule-build-perl fail to work
properly.
(From OE-Core rev: 2d5fae94a889cd9558e3946ecec2362c94e1c1e9)
Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit e4072d53a7fb4fbbdacce9a20968e71ef6cff307)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Drop fix-setgroup.patch as the upstream has fixed the issue.
(From OE-Core rev: df86f5d32dc18f25c5c4788b65cceae8905a6ef3)
Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 45edc6d23e20f7634c50db71e419c7e3bb7f393e)
[Bug fix only update]
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Add a patch which handles the following issues:
a) Remove the \n from configure_attr.sh since it gets quoted differently depending on
whether the shell is bash or dash which can cause the test result to be incorrect.
Reported upstream: https://github.com/arsv/perl-cross/issues/87
b) Sort the order of the module lists from configure_mods.sh since otherwise
the result isn't the same leading to makefile differences.
Reported upstream: https://github.com/arsv/perl-cross/issues/88
c) Sort the Encode::Byte byte_t.fnm file output (and the makefile depends whilst
there for good measure)
This needs to go to upstream perl (not done)
d) Use bash for perl-cross configure since otherwise trnl gets set to "\n" with bash
and "" with dash
Reported upstream: https://github.com/arsv/perl-cross/issues/87
(From OE-Core rev: 482fd0d99f989b5a72a25bdf402fb2f219420b5d)
(From OE-Core rev: def3a9d748564883d71c506726554df622701b00)
(From OE-Core rev: 1f630fe43ec3c3e78c25f93d6badc8a35ff782ad)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There is a bug in patch 2.7.3 and earlier where index lines
in patches can change file modes when they shouldn't:
http://git.savannah.gnu.org/cgit/patch.git/patch/?id=82b800c9552a088a241457948219d25ce0a407a4
This leaks into debug sources in particular (e.g. tcp-wrappers where
source files are read-only). Add the dependency to target recipes
to avoid this problem until we can rely on 2.7.4 or later.
We could try and remove all index lines from patch files but it will be a
losing battle. We could try and identify all the recipes which change
modes on files in patches but again, its a losing battle.
Instead, compromise and have patch-native as a dependency
for target recipes. We use patch-replacement-native since patch-native
is in ASSUME_PROVIDED.
Also add nativesdk-patch to buildtools-tarball.
[YOCTO #13777]
(From OE-Core rev: 5ed0840c93804488cd1c1aba6cb382b2434714a5)
(From OE-Core rev: fd3bd61a6fe5190c575dc968f3a0be9c1cbf21ed)
(From OE-Core rev: 148f1f8caf5d9a262c1f55e437326ce6139a743e)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There is a sorting problem with opkg-build where the ipk generated is depending
upon the order of files on disk. The reason is the --sort option to tar only
influences the orders of files tar reads, not those passed by the -T option.
Add in a sort call to resolve this issue. To ensure consistent sorting we
also need to force to a specific locale (C) else the results are still not
deterministic.
(From OE-Core rev: a9b8287984c63420e10329a69f7ac5125f1687f8)
(From OE-Core rev: b577a6d923042cfc04e67d470e0987488ea61412)
(From OE-Core rev: ff31fa7ae18cffb1618c3859c5dff7eb3c587692)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Linux kernel compilation for aarch64 triggers ICE if
CONFIG_CC_OPTIMIZE_FOR_SIZE=y.
The rootcause is GCC bug #91102 'aarch64 ICE on Linux kernel with -Os'.
Apply the fix to 9.2.
(From OE-Core rev: 14f34d32bfdaa752f5043e62750d2e7b92c4b419)
(From OE-Core rev: 8ebd3b4ed4995f27c1568cf873067ce24b1998bd)
Signed-off-by: Taras Kondratiuk <takondra@cisco.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Linux system's realpath() remove trailing slashes, but pseudo's doesn't, need
make them identical.
E.g., the following code (rel.c) prints '/tmp' with system's realpath, but
pseudo's realpath prints '/tmp/':
#include <stdio.h>
#include <limits.h>
#include <stdlib.h>
int main() {
char out[PATH_MAX];
printf("%s\n", realpath("/tmp/", out));
return 0;
}
$ bitbake base-passwd -cdevshell # For pseudo env
$ gcc rel.c
$ ./a.out
/tmp/ (but should be /tmp)
This patch fixes the problem.
(From OE-Core rev: c251e753486dae9b460df315a6d19d6c81098ecb)
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 319bbf66e03377adf2db7efa93ef578e3460eb38)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Update to account for review feedback on list.
(From OE-Core rev: 9a2748db44c4382bbba81a5a9b96c998f0fab983)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit ec788594f3f6a47687c6eb321437f2d2b58b1518)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This is not exploitable when glibc has CVE-2016-10739 fixed,
which is fixed in the upstream version since warrior.
(From OE-Core rev: a26ac2921a1ad96959364223920402082ccd1d61)
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
One Windows-only CVE that cannot be fixed, and two CVEs
where upstream agreement is that they are not vulnerabilities.
(From OE-Core rev: 56d5b181f3b119f2bbd310dedd6d3b26e76f5944)
(From OE-Core rev: 13024049625c1705108066b38396ac379aacce84)
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Chee Yang Lee
Changqing Li
Adrian Bunk
haiqing
Li Zhou
Jan Luebbe
Khem Raj
Jeremy Puhlman
Richard Purdie
Anuj Mittal
Rahul Chauhan
Nathan Rossi
Mark Hatle
Tim Orling
Alexander Kanavin
Ross Burton
Alejandro del Castillo
Taras Kondratiuk via Openembedded-core
Armin Kuster
Robert Yang
Joshua Watt
Stefan Müller-Klieser
Peter Kjellerstedt
Christopher Larson
Hongxu Jia
Vinay Kumar