rpmbuild can start processing random memory when processing the value
provided by XZ_THREADS, and unintentionally disable encoding for a
file descriptor that in fact requires encoding to be enabled in order
for lzwrite() to actually create an rpm.
*** Fdopen(0x7f2030002b30,w6T16.xzdio) | fdio 23 fp (nil)
==> lzopen_internal("w6T16", 23, 1)
==> lzopen_internal set encoding
==> lzopen_internal clear encoding
==> Fdopen(0x7f2030002b30,"w6T16.xzdio") returns fd 0x7f2030002b30 | xzdio 0x7f2030004e30 fp 23 | fdio -1 fp (nil)
==> lzwrite(0x7f2030004e30, 0x7f20789d8070, 6) encoding 0
==> Fwrite(0x7f2030002b30,0x7f20789d8070,6) rc -1 | xzdio 0x7f2030004e30 fp 23 | fdio -1 fp (nil)
error: create archive failed: cpio: write
When the encoding bit gets cleared on the LZFILE* struct, lzwrite() then
rightfully complains when it detects !lzfile->encoding, which then gets
bubbled up as a write failure when we go to create the archive.
This fix is available in the rpm 4.17-release.
(From OE-Core rev: b093005d31467d89b00af621f86eb5cac7f845af)
Signed-off-by: Kyle Russell <bkylerussell@gmail.com>
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This update was made with the convert-scruri.py script in scripts/contrib
This script handles two emerging issues:
1. There is uncertainty about the default branch name in git going forward.
To try and cover the different possible outcomes, add branch names to all
git:// and gitsm:// SRC_URI entries.
2. Github are dropping support for git:// protocol fetching, so remap github
urls as needed. For more details see:
https://github.blog/2021-09-01-improving-git-protocol-security-github/
(From OE-Core rev: 827a805349f9732b2a5fa9184dc7922af36de327)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
On an aarch64 build host, vendor is found to be "unknown", on x86 systems
it is "pc". This filters through to the PLATFORM tag in target rpms.
We saw reproducibility test failures where the PLATFORM tags in noarch
rpms were changing depending upon which host built them. Forcing the
vendor value to a consistent one makes things deterministic.
(From OE-Core rev: b7dfe230b9b40145f43fa0bd42be82ae41a3ef3e)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit f6434075b2bdfc23c683d22281b674b1e6abde77)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
0001-rpm-rpmio.c-restrict-virtual-memory-usage-if-limit-s.patch changed
to avoid critical issues
Handled return values of getrlimit() and lzma_cputhreads() functions
to avoid unexpected behaviours like devide by zero and potential read
of uninitialized variable 'virtual_memory'
Upstream-Status: Pending [merge of multithreading patches to upstream]
(From OE-Core rev: ad080aadbc409c99511d602e0531952b96c06bbf)
Signed-off-by: Ranjitsinh Rathod <ranjitsinh.rathod@kpit.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 5aae9c2cb464350bc443a0f60fd6602942e61f46)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Adding fix for CVE-2021-20266
Upstream-Status: Backport [9646711891]
Note: Hunk#2 and Hunk#3 refreshed to apply patch and match value of
dl_max variable to make it with current version
All Hunks are refreshed to solve patch-fuzz
(From OE-Core rev: 6c16aad7167eb98bc9995486f967431c39f9df15)
Signed-off-by: Ranjitsinh Rathod <ranjitsinh.rathod@kpit.com>
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
unsigned signature header leads to string injection into an rpm database
reference:
https://nvd.nist.gov/vuln/detail/CVE-2021-3421d6a86b5e69
(From OE-Core rev: eb1183cac91a9948ac47acc56454b8841764ba72)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
For now, the nativesdk-rpm's %_var in /usr/lib/rpm/macros is
'/var'. This is causing error when running `rpm -qplv A.rpm'.
error: cannot open Packages index using db5 - Permission denied (13)
error: cannot open Packages database in /var/lib/rpm
The rpm in SDK should be using its own database. So we should remove
this configure option, letting the SDK's native sysroot prefix to be
there in %_var. In fact, '%_usr' in macros has already got the prefix.
After this change, we have in the macros file lines like below.
"""
%_usr /opt/windriver/wrlinux/20.29/sysroots/x86_64-wrlinuxsdk-linux/usr
%_usrsrc %{_usr}/src
%_var /opt/windriver/wrlinux/20.29/sysroots/x86_64-wrlinuxsdk-linux/var
"""
(From OE-Core rev: 4e0d1bdc6e896c7dedccf615d6d0da9eb91d9b1d)
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 13e17930062cb3f816516ba7dbeb70d6da7174dd)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
nss is a problematic library in many ways, but openssl
seems to work just fine.
rpm was the only user of nss in oe-core, which opens
the possibility of moving it out.
Add a backported patch to enable builds with latest openssl.
(From OE-Core rev: 8c35a01dc771bf48f74129daa58d365100d18081)
Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This is perhaps not the most elegant patch but it be dropped
once rpm is updated to the soon-to-be-released version 4.15
where upstream has removed mono dependency resolution altogether.
(From OE-Core rev: c829e89f99c973787f31422b4c32eeebe8c383d3)
Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
A couple have still been missed in the past despite multiple
attempts at doing so (or simply have re-appeared?).
Search & replace made using the following command:
sed -e 's|\(d\.getVar \?\)( \?\([^,()]*\), \?True)|\1(\2)|g' \
-i $(git grep -E 'getVar ?\( ?([^,()]*), ?True\)' \
| cut -d':' -f1 \
| sort -u)
(From OE-Core rev: 9f551d588693328e4d99d33be94f26684eafcaba)
Signed-off-by: André Draszik <andre.draszik@jci.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Previously there was lots of irrelevant noise in the logs,
because we also wanted to be able to debug postinst issues
easily. I have adjusted the logging levels so that
postinst info is still written to the logs, but other
things are not.
[YOCTO #13119]
(From OE-Core rev: ffb7b8f70937a7d95814c1a99527d5ea7cbf7cee)
Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Fix the situation where the task do_package_write_rpm ends up in
"liblzma: memory allocation failed". This happens if the host
environment has set a limit on virtual_memory for the user with
'ulimit -v' for packages with a lot of binary packages, e.g. glibc-locale.
(From OE-Core rev: a937cff2746073d1dea37d85e7305d8d6705ff28)
Signed-off-by: Peter Bergin <peter@berginkonsult.se>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Drop --disable-dbus option and patch: --disable-plugins is sufficient.
Drop upstreamed patches.
Drop a chunk of 0001-Do-not-reset-the-PATH-environment-variable-before-ru.patch, as upstream fixed the issue.
(From OE-Core rev: 4e4fc46589762914f10ae2fa41b614bcb32f9bb9)
Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
RPM writes each package scriptlet (post-/preinstall) to
/var/tmp/rpm-tmp.XXXXXX --- a lot of files potentially gets created.
When debugging is enabled, these temporary scriptlet files aren't
cleaned up at all and after a while this results in the filesystem
resources are eaten up (like running out of available inodes).
Normally, the temporary files would have been written to the tmp
directory of the target sysroot (which we can easily clean up), but in
this tree, you can't necessarily run the scriptlets.
Fixes [YOCTO #12792]
(From OE-Core rev: ffb0ece83e74797f4c3da3866bb3d691c388a5e5)
Signed-off-by: Olof Johansson <olofjn@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Recent patches were applied which removed a function which set the xx
variable. This means xx can be uninitalised and prerm functions can fail.
This adjusts the patch to ensure the xx value isn't used in such a way.
(From OE-Core rev: 23f15c63777020f5d43b070a1eb2bcf246c19ff8)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
If the maximum number of open file descriptors is much greater than the
usual 1024 (for example inside a Docker container), the performance
drops significantly.
This was reported upstream in:
https://bugzilla.redhat.com/show_bug.cgi?id=1537564
which resulted in:
https://github.com/rpm-software-management/rpm/pull/444
The pull request above has now been integrated and this commit contains
a backport of its three patches, which together change the behavior of
rpm so that its performance is now independent of the maximum number of
open file descriptors.
(From OE-Core rev: 7feed9ccfc4e656c6264f07e13d7e9ef69bdfb06)
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This reverts commit 6f1822e5f1eaafd8bc46e999de730c1fcca77f3a.
This patch only solved a part of the problem.
(From OE-Core rev: 78c4eb6ea4230a4f9199aa3fa7b84d2aea465b95)
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When installing the SDK to a non-default path, running "rpm --showrc" from the
sdk will produce the following error:
error: Unable to open /opt/windriver/wrlinux-small/10.17.41/sysroots/x86_64-wrlinuxsdk-linux/usr/lib/rpm/rpmrc for reading: No such file or directory.
Fix this by adding wrappers that dynamically export the RPM_CONFIGDIR,
RPM_ETCCONFIGDIR and MAGIC environment variables, pointing to the proper
sdk locations.
(From OE-Core rev: 760103cdaed3e820888d8984ec0b76cfc831d534)
Signed-off-by: Ovidiu Panait <ovidiu.panait@windriver.com>
Signed-off-by: Yi Zhao <yi.zhao@windriver.com>
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Move the recipe specific over-ride for ASNEEDED into the recipe to
make it more apparent that the over-ride is being applied (and that
it should be re-checked on version updates, etc).
(From OE-Core rev: f3d223304e52b9be946e5bd849075147147cbbb3)
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>
Add option for dbus in configure.ac, and explicitly build without dbus
for rpm-native.
Previously, the rpm recipe tries to prevent rpm-native from attempting
to inhibit shutdown via session dbus by appending '--disable-plugins'
to EXTRA_OECONF in case of native.
However, some layer may need some functionality via plugin support. And
when it enables it, we would meet the following warning at rootfs time.
Unable to get systemd shutdown inhibition lock: Socket name too long
As plugins/systemd_inhibit.c is the only place that's related to this
dependency, we can be sure that dbus is really not needed for rpm-native.
(From OE-Core rev: 815ccef2d5bef1e46c51916e694d0974aee394a9)
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Drop two upstreamed patches.
License-Update: grammar fixes
(From OE-Core rev: c74bbe70988e7dd3ee1b7977de68a7ffe64fc8ef)
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Backport upstream configure fix for hosts that have multiple users with UID
0 or groups with GID 0.
(From OE-Core rev: f0f92d3d0f42a4a7e521b58dac53e14f9e2572a1)
Signed-off-by: Mike Crowe <mac@mcrowe.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Previously oe-core had a development snapshot of rpm, it's better to update
to something more stable.
Removed patches:
0001-Add-PYTHON_ABI-when-searching-for-python-libraries.patch
(upstream is using pkg-config)
0001-When-nice-value-cannot-be-reset-issue-a-notice-inste.patch
(functionality has been moved to a plugin, we disable plugins
for rpm-native)
0012-Use-conditional-to-access-_docdir-in-macros.in.patch
(merged upstream)
Changed patches:
0001-Fix-build-with-musl-C-library.patch
(one previous musl issue has been resolved upstream; another has been added)
Rest of the patches are trivial rebases.
Update the signing oe-selftest so that the reference output matches
the upstream changes.
(From OE-Core rev: b4613b6ce07c295c5d6de6861acf19315acaccb2)
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We do not want to specify --sysroot when defining __cc
used on a target.
(From OE-Core rev: 0ff69fc42b0ac2e7e865943e7febd87093da69d3)
Signed-off-by: Joe Slater <jslater@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When rpmdeps files a perl script, it attempts to determine what it provides
and what it requires. Often the requires are incorrect, within the context
of Wind River Linux. This results in an error that DNF is unable to install
a package due to one or more unresolved dependencies.
In RPM5 we had disabled this behavior, the alternative is to require that all
perl scripts be 'complete', in that they only require things they absolutely
need and that OE provides. If we ever enforce that, this commit can be
reverted. Until they fall back to prior behavior (which also matches ipkg
and deb style packages.)
(From OE-Core rev: 99376f602c997fec96dbb28ddbe3fa5f68fde2c3)
(From OE-Core rev: 586875cea5769fea7d17bae2677eab72e4fc7e31)
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Also modify a Python script (pythondistdeps.py) to use Python 3.
(From OE-Core rev: b2906dbebfd547a630a4eafdcbaa1ceefa5e257c)
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
... instead of a global exception list which was problematic.
[YOCTO #11896]
(From OE-Core rev: 89dfede4ca795ba085f1ee7290c6dede573c11db)
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
RPM file signing is enabled with --with-imaevm during configuration.
If enabled, the RPM signing tool rpmsign will call libimaevm.so provided
by the recipe ima-evm-utils.
(From OE-Core rev: 9d7797e4374c111f0dff523fd49354bcc33dc2af)
Signed-off-by: Lans Zhang <jia.zhang@windriver.com>
Signed-off-by: Jia Zhang <lans.zhang2008@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This fixes a following error:
| recipe-sysroot-native/usr/lib/rpm/debugedit: error while loading shared libraries: \
| libbz2.so.1: cannot open shared object file: No such file or directory
(From OE-Core rev: 61ecda8f7977ee2d30f31d2f384f65f933971568)
Signed-off-by: Ming Liu <peter.x.liu@external.atlascopco.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This greatly reduces build times when there is a large amount of small
rpm packages to produce. The patches are rather invasive,
and so will be submitted upstream.
(From OE-Core rev: 964a6eb4732df462008883c4bb003f801777dfad)
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This will send the output from rpmfcPrint() to stdout. This is an
alternative to using the --rpmfcdebug option, which will send the same
output to stderr. The two options have totally different use cases
though. While --alldeps is used when the output from rpmfcPrint() is
what is wanted, --rpmfcdebug can be used together with the other
output options, e.g., --requires, without affecting their output.
(From OE-Core rev: a58a0dc03398dcd9f81a9c8a6189ae13d90e0df7)
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This avoids the following warning:
warning: Ignoring invalid regex %{_docdir}
when runing `rpmdeps -R <file>`, since %{_docdir} is only defined when
parsing a spec file.
(From OE-Core rev: 7e0964e506506d20a25aac570104938759f9f70e)
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There is nothing that requires, e.g., a DSO to be executable, but it
is still an ELF binary and should be identified as such.
(From OE-Core rev: 74d040f456269992a13850a626279b3b8e954847)
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Rather than trying to call rpmdeps with the correct arguments to work
with the sysroot as was done in package.bbclass, create a wrapper for
it like all the other native tools already had.
(From OE-Core rev: 124a6115af845fd892f53c8504db6ffd59f8bd45)
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Use a loop rather than calling create_wrapper for each individual
tool.
(From OE-Core rev: e299a396849acd515d51d377958d795e979e6262)
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Currently the file encodes full paths to various host tools in the
HOSTTOOLS directory which is bad in native and target cases. We can
simply use the versions from PATH quite safely in OE.
(From OE-Core rev: be901200d94beaa35e1d05eb502b117b3b523609)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
These tools are not currently used for anything, but we should
still provide working versions of them.
[YOCTO #11400]
(From OE-Core rev: da11fbde6f8164e2774068b99bab84e4b6084713)
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The rpm 5 has a rpm-build package, so here should use RPROVIDES rather
than PROVIDES to keep compatibility.
(From OE-Core rev: de2ee88f9cc0fc8d6d92ac2a79364e79a99ae98e)
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
rpm 5.x was packaging build tools separately, so we need to unbreak
things that relied on that.
[YOCTO #11167]
(From OE-Core rev: 3b5ac72bdf76ac8ff98dc3c882a4edc77c6e2c33)
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The dnf stack is written and tested against rpm 4.x. So if we want to use dnf for packaging,
we should also use rpm 4 - there's simply too much work involved in making rpm 5 work with it due
to significant API differences, and supporting that going forward.
(From OE-Core rev: 2358e786ec8d1199d90e181eb5d8d00816f669b4)
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The source code is incompatible with rpm4 API - let's use rpm
binary itself for now.
(From OE-Core rev: 127112fa05c7102dacda4173adf380c557bacecb)
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
