]> code.ossystems Code Review - openembedded-core.git/log
openembedded-core.git
3 years agowic: support rootdev identified by partition label
Schmidt, Adriaan [Tue, 23 Nov 2021 08:33:29 +0000 (08:33 +0000)]
wic: support rootdev identified by partition label

We already support specifying the rootfs by PARTUUID. This adds general
support for letting the kernel find the rootfs by PARTLABEL.

Signed-off-by: Adriaan Schmidt <adriaan.schmidt@siemens.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agobitbake.conf: Pass -D option to ranlib for determisim
Richard Purdie [Mon, 22 Nov 2021 16:08:16 +0000 (16:08 +0000)]
bitbake.conf: Pass -D option to ranlib for determisim

Add the -D option to BUILD_RANLIB so that deterministic archives
are built for native/cross output. This improves the changes of hash
equivalence matches and hence build artefact reuse.

We don't need this in the target case since we compile binutils-cross
with an option making this the default.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agobuildhistory: Fix do_package race issues
Richard Purdie [Tue, 23 Nov 2021 13:14:35 +0000 (13:14 +0000)]
buildhistory: Fix do_package race issues

The buildhistory_list_pkg_files function uses data from do_package, not
do_packagedata. Usally the two are restored together but it may see
a half complete directory or other races issues depending on timing.

Rework the function so that it uses the correct task dependencies. This
should avoid races but means the data is only restored to buildhistory
if the do_package or do_package_setscene tasks are restored.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agoglibc-tests: Do not set SYSTEMD_SERVICE
Khem Raj [Thu, 18 Nov 2021 22:43:01 +0000 (14:43 -0800)]
glibc-tests: Do not set SYSTEMD_SERVICE

These are glibc specific which comes from glibc packaging class

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agoptest-packagelists: Handle glibc vs musl for ptest
Richard Purdie [Thu, 11 Nov 2021 17:22:41 +0000 (17:22 +0000)]
ptest-packagelists: Handle glibc vs musl for ptest

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agoglibc: ptest: Add running glibc time related test suite (Y2038) with ptest
Lukasz Majewski [Tue, 9 Nov 2021 15:28:51 +0000 (16:28 +0100)]
glibc: ptest: Add running glibc time related test suite (Y2038) with ptest

This patch introduces new recipe - namely 'glibc-tests', which
builds and installs time related (to check if Y2038 support works) glibc
test suite to OE/Yocto built image.

It reuses code from already available 'glibc-testsuite' recipe,
which is run with 'bitbake glibc-testsuite -c check' and uses qemu
to execute remotely (via SSH) tests on some emulated machine.

This recipe installs time related glibc tests on some rootfs image.
Afterwards, those tests can be executed on the real hardware, to
facilitate validation of it with Y2038 problem compliance.

To test time related subset - one needs to call:
ptest-runner glibc-tests
then change the date after Y2038 threshold for 32 bit systems:
date -s "20 JAN 2038 18:00:00"
and then run ptest-runner again.

To facilitate debugging, source files are provided by default with
the unstripped debugging symbols. Such approach would reduce the
already complex recipe (as it inherits base glibc one), so there
is no need to also install *-dbg and *-src packages.

Signed-off-by: Lukasz Majewski <lukma@denx.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agogmp: fix CVE-2021-43618
Ross Burton [Mon, 22 Nov 2021 11:40:56 +0000 (11:40 +0000)]
gmp: fix CVE-2021-43618

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agolibjpeg-turbo: upgrade 2.1.1 -> 2.1.2
wangmy [Mon, 22 Nov 2021 14:58:31 +0000 (22:58 +0800)]
libjpeg-turbo: upgrade 2.1.1 -> 2.1.2

Significant changes relative to 2.1.1:
Fixed a regression introduced by 2.1 beta1[13] that caused the remaining
GAS implementations of AArch64 (Arm 64-bit) Neon SIMD functions
(which are used by default with GCC for performance reasons) to be placed in
the .rodata section rather than in the .text section.
This caused the GNU linker to automatically place the .rodata section in
an executable segment, which prevented libjpeg-turbo from working properly
with other linkers and also represented a potential security risk.

Fixed an issue whereby the tjTransform() function incorrectly computed the
MCU block size for 4:4:4 JPEG images with non-unary sampling factors and
thus unduly rejected some cropping regions, even though those regions aligned
with 8x8 MCU block boundaries.

Fixed a regression introduced by 2.1 beta1[13] that caused the build system
to enable the Arm Neon SIMD extensions when targetting Armv6 and other legacy
architectures that do not support Neon instructions.

libjpeg-turbo now performs run-time detection of AltiVec instructions on
FreeBSD/PowerPC systems if AltiVec instructions are not enabled at compile time.
This allows both AltiVec-equipped and non-AltiVec-equipped CPUs to be supported
using the same build of libjpeg-turbo.

cjpeg now accepts a -strict argument similar to that of djpeg and jpegtran,
which causes the compressor to abort if an LZW-compressed GIF input image
contains incomplete or corrupt image data.

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agopython3-mako: upgrade 1.1.5 -> 1.1.6
wangmy [Mon, 22 Nov 2021 14:58:33 +0000 (22:58 +0800)]
python3-mako: upgrade 1.1.5 -> 1.1.6

Fixed issue where control statements on multi lines with a backslash would
not parse correctly if the template itself contained CR/LF pairs as on
Windows.

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agomesa: upgrade 21.2.5 -> 21.3.0
wangmy [Mon, 22 Nov 2021 14:58:32 +0000 (22:58 +0800)]
mesa: upgrade 21.2.5 -> 21.3.0

New features:
VK_EXT_color_write_enable on lavapipe
GL_ARB_texture_filter_anisotropic in llvmpipe
Anisotropic texture filtering in lavapipe
VK_EXT_shader_atomic_float2 on Intel and RADV.
VK_EXT_vertex_input_dynamic_state on RADV.
VK_KHR_timeline_semaphore on lavapipe
VK_EXT_external_memory_host on lavapipe
GL_AMD_pinned_memory on llvmpipe
GL 4.5 compatibility on llvmpipe
VK_EXT_primitive_topology_list_restart on RADV and lavapipe.
ES 3.2 on zink
VK_KHR_depth_stencil_resolve on lavapipe
VK_KHR_shader_integer_dot_product on RADV.
OpenGL FP16 support on llvmpipe
VK_KHR_shader_float16_int8 on lavapipe
VK_KHR_spirv_1_4 on lavapipe
Experimental raytracing support on RADV
VK_KHR_synchronization2 on Intel
NGG shader based culling is now enabled by default on GFX10.3 on RADV.
VK_KHR_maintenance4 on RADV
VK_KHR_format_feature_flags2 on RADV.
EGL_EXT_present_opaque on wayland

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agolibcap: upgrade 2.60 -> 2.61
wangmy [Mon, 22 Nov 2021 14:58:30 +0000 (22:58 +0800)]
libcap: upgrade 2.60 -> 2.61

better number parsing in capsh and setcap;
fixed segfaulting in .so executables;
added example of capable shared library object.

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agopython3-scons: upgrade 4.2.0 -> 4.3.0
wangmy [Mon, 22 Nov 2021 14:59:34 +0000 (22:59 +0800)]
python3-scons: upgrade 4.2.0 -> 4.3.0

NOTE: 4.3.0 now requires Python 3.6.0 and above. Python 3.5.x is no longer supported

Changelog:
- Default hash algorithm check updated for SCons FIPS compliance. Now checks for hash viability
  first and then walks the tree to use the first viable hash as the default one. This typically
  selects SHA1 on FIPS-enabled systems less than Python 3.9 as the new default instead of MD5,
  unless SHA1 has also been disabled by security policy, at which point SCons selects SHA256
  as the default. For systems running Python 3.9 and later, the hashlib bug has been fixed,
  and SCons will once again default to MD5 as the preferred algorithm.
- Fix MSVS tests (vs-N.N-exec.py) for MSVS 6.0, 7.0, and 7.1 (import missing module).
- Add support for Visual Studio 2022.
- Fix reproducible builds. Restore logic respecting SOURCE_DATE_EPOCH when set.
- Fix version tests to work with updated scons --version output. (Date format changed)
- Fix issue #4021.  Change the way subst() is used in Textfile() to not evaluate '$$(' -> '$',
  but instead it should yield '$('.
- Change SCons.Platform.win32.get_architecture() to return platform.platform() when run in an
  environment where neither: PROCESSOR_ARCHITEW6432 nor PROCESSOR_ARCHITECTURE is set.
  This should fix platform tests which started failing when HOST_OS/HOST_ARCH changes
  introduced by Aaron Franke (listed below) were merged.
- Further PCH updates. It's now recommended that env['PCH'] should always be a File node.
  Either via return value from env.PCH() or by explicitly using File('StdAfx.pch').
- Added --no-ignore-skips to runtest.py. Changed default to ignore skips when setting
  runtest.py's exit status. Previously would exit 2 if any tests were skipped.
  Now will only exit 2 if user specifies --no-ignore-skips and some tests were skipped.
- Small fix to ensure CLVar default value is an empty list.
  See MongoDB bug report: https://jira.mongodb.org/browse/SERVER-59656
  Code contributed by MongoDB.
- Ninja - Fixed an issue where if you control-c and/or killed ninja while it was running scons to
  regenerate build.ninja you would end up with no build.ninja file and have to rerun scons from scratch.
  Code contributed by MongoDB.
- Define HOST_OS and HOST_ARCH in the environment for all platforms.
  Before this change, these were only defined for Win32 and OS/2.
- Fix ninja tool to never use for_sig substitution because ninja does not use signatures. This
  issue affected CommandGeneratorAction function actions specifically.
- Expanded ninja Mkdir to also support Mkdir actions.
- Added support for the PCH environment variable to support subst generators.
- Fix command line escaping for ninja dollar sign escape. Without escaping ninja properly,
  the ninja file scons regenerate and callback invocations will lose the $ characters used in
  the scons command line which ninja uses itself for escaping. For Example:
      scons BUILD=xyz OTHERVAR=$BUILD
  Prior to this fix, it would cause ninja to fail to escape the dollar sign, leading to the
  single dollar sign being used as a ninja escape character in the ninja file.
- Added ninja API 'NINJA_FORCE_SCONS_BUILD' to force a node to callback to scons.
- Two small Python 3.10 fixes: one more docstring turned into raw
  because it contained an escape; updated "helpful" syntax error message
  from 3.10 was not expected by SubstTests.py and test/Subst/Syntax.py
- EmitterProxy rich comparison set is completed (checker warning).
  Added __le__, __gt__, __ge__.
- Fix gcc/g++ tool failing if "gcc --version" returns text which fails
  to_String conversion (i.e., not UTF-8) - failure happens when tool
  initialization checks version. For gcc, the initial version string is
  not translated, for the rest, don't convert, just consume raw and discard.
- Maintenance and doc: modernize some usage in Scanner package,
  calling super(), switching some imitialization to comprehensions,
  and code formatting.  Docstring for scanner Base moved from
  init-method to class-level so it's picked up by Sphinx.
  Added new sconsign filenames to skip_entry_list in Scanner/Dir.py
- Change SCons.Scanner.Base to ScannerBase. Old name kept as an alias
  but is now unused in SCons itself.
- Call Variables option converter consistently - the converter should
  have access to the env if it needs to (issue #2064).
- Fixed the variables Add() method to accept a tuple for the variable
  name the same way AddVariables() does (issue #3869).
- The premade validator PathIsDirCreate for for PathVariable now catches
  the case where the directory could not be created due to permission
  problems, allowing a more helpful error to be emitted (issue #2828)
- Maintenance: Python thread.setDaemon is deprecated in favor of
  directly updating daemon attribute - update SCons to do this.
- Make sure when subst'ing a callable, the callable is called with
  the correct for_signature value, previously it would be true even
  if doing SUBST_RAW (issue #4037)
- Update Util/NodeList implementation to get rid of a workaround for
  early Python 3 slicing issue that is no longer a problem.
- Rework some Java tests to skip rather than fail on CI systems, where
  the working java is > v9, but a 1.8 or 9 was also found.
- Java updates: on Windows, detect more default JDK install locations.
  On all platforms, more Java versions (up to 17.0 now).  Add more information
  on version selection to docs.
  Update docs on JavaH tool in light of javah command dropped since 10.0.
  Try to be better about preserving user's passed-in JAVA* construction vars.
- Start the deprecation of the qt tool, which refers to Qt3 (usupported
  since around 2006). There's a deprecation warning added, initially
  defaulting to disabled.
- Fix crash when scons is run from a python environement where a signal
  is set from outside Python.

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agosystemtap: upgrade 4.5 -> 4.6
wangmy [Mon, 22 Nov 2021 14:59:33 +0000 (22:59 +0800)]
systemtap: upgrade 4.5 -> 4.6

support for 64-bit RISC-V architecture, liveness analysis for guru-mode write operations,
bpf syscall_any and abort() tapsets, bpf foreach iteration of multi-key arrays,
return of inter-cpu output ordering

Changelog is as follows:
= SystemTap frontend (stap) changes
- stap-prep now tries to download the main kernel debuginfo file from
  a debuginfod server, if configured.

= SystemTap backend changes
- SystemTap has added support for the 64-bit RISC-V architecture.
- SystemTap now uses DynInst to perform a liveness analysis on
  target variables and warn when a guru-mode modification to a variable
  will have no effect. The liveness analysis is currently done on
  x86_64, PowerPC, and AArch64.
- The kernel-user relayfs transport again sorts messages into a total
  time order across CPUs.  High output-volume scripts may need a
  larger "-s BUF" parameter to reliably transfer.  "-b" bulk mode
  is also available again as an alternative.
- The bpf backend now supports foreach iteration in multi-key associative arrays.

= SystemTap tapset changes
- Updated syscall_any tapset mapping to include newer syscalls.
- syscall_any tapset can be used by the bpf backend.
- abort() tapset can be used by the bpf backend.

= Known issues with this release
- There are known issues on kernel 5.10+ after adapting to set_fs()
  removal, with some memory accesses that previously returned valid data
  instead returning -EFAULT (see PR26811).
- An sdt probe cannot parse a parameter that uses a segment register.
  (PR13429)
- The presence of a line such as
      *CFLAGS += $(call cc-option, -fno-var-tracking-assignments)
  in older linux kernel Makefile unnecessarily reduces debuginfo quality,
  consider removing that line if you build kernels.  Linux 5.10+ fixes this.

= Bugs fixed for this release <https://sourceware.org/PR#####>
6562     $SYSTEMTAP_DEBUGINFO_PATH does not work
15724     stapdyn looking for libdyninstAPI_RT.a
26839     Systemtap build failures with clang
27820     abort() tapset not implemented in the bpf mode
27829     support for floating point values passed through sdt.h markers
27864     loc2stap.cxx assertion failure on loc_unavailable type location, rawhide
27881     failed to extend vma mapped entry when the address is adjacent
27903     handle f33 glibc $$parms
27932     List Python as a prerequisite in README
27933     Use of unitialized functioncall synthetic field in
27934     failure to attach statement
27940     The /* pc=0x... */ is no longer printed by "stap -v -L 'kernel.function("*")'
27942     testsuite/systemtap.base/perf.sh drop bashism
27984     stap skipping partially-inlined instance, but it is not inline function actually
28070     extend vma end address to the different module
28079     adapt to kernel 5.14 task_struct.__state change
28084     autoconf-x86-uniregs.c compile failled with -Werror cause STAPCONF_X86_UNIREGS missing
28140     kernel panic on tracepoint activation in stap module
28184     task_fd_lookup failed on linux 5.11
28244     linux objtool imposes symbol length limits on generated function names
28384     finish nfs_proc tapset port 4.3 string server_ip
28443     Provide syscall_any tapset for bpf
28449     loss of cross-cpu output ordering
28544     procfs_bpf.exp regression due to string handling error
28557     module kprobe insertion on modern kernels

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agostress-ng: upgrade 0.13.06 -> 0.13.07
wangmy [Mon, 22 Nov 2021 14:59:32 +0000 (22:59 +0800)]
stress-ng: upgrade 0.13.06 -> 0.13.07

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agobuildhistory: Fix srcrevs output
Richard Purdie [Mon, 22 Nov 2021 15:32:36 +0000 (15:32 +0000)]
buildhistory: Fix srcrevs output

The code was assuming that the a recipe with only one srcrev wouldn't "name"
it. This isn't the case as the glibc or bzip2 recipes show, you can have
a single srcrev which is named.

We can pull the data from the fetcher and in fact we already have it, we just
need to handle the "default" case and make that code the default for all srcrev
regardless of length.

[YOCTO #14017]

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agobuildhistory: Drop support for older bitbakes
Richard Purdie [Mon, 22 Nov 2021 15:21:49 +0000 (15:21 +0000)]
buildhistory: Drop support for older bitbakes

We've bumped the minimum bitbake version past the point this fallback code
was needed, drop it.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agoclasses/crate-fetch: Ensure crate fetcher is available
Joshua Watt [Fri, 19 Nov 2021 02:12:49 +0000 (20:12 -0600)]
classes/crate-fetch: Ensure crate fetcher is available

Reworks the crate fetcher class to have it install the fetcher at recipe
finalization so that it is always available before SRC_URI is expanded.
In addition, override the value of SRCPV to also install the fetcher
when SRCPV is expanded so that AUTOREV works.

[YOCTO #10867]

Signed-off-by: Joshua Watt <JPEWhacker@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agorootfs-postcommands: update systemd_create_users
Vyacheslav Yurkov [Fri, 19 Nov 2021 21:05:39 +0000 (22:05 +0100)]
rootfs-postcommands: update systemd_create_users

Process all systemd-sysusers configuration files in order to create
users/groups at build time. systemd-sysusers would try to create them at
run-time, but for read-only rootfs that's not possible and results in
warnings from different services:

systemd-udevd[166]: /lib/udev/rules.d/50-udev-default.rules:39 Unknown group 'render', ignoring
systemd-udevd[166]: /lib/udev/rules.d/50-udev-default.rules:40 Unknown group 'render', ignoring
systemd-udevd[166]: /lib/udev/rules.d/50-udev-default.rules:42 Unknown group 'sgx', ignoring

Signed-off-by: Vyacheslav Yurkov <uvv.mail@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agoclasses/meson: Add optional rust definitions
Joshua Watt [Fri, 19 Nov 2021 20:51:40 +0000 (14:51 -0600)]
classes/meson: Add optional rust definitions

Adds the rust tools to the cross and native files if present so that
projects that use both rust and meson can build

Signed-off-by: Joshua Watt <JPEWhacker@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agooeqa: fix warnings for append operators combined with +=
Yi Zhao [Fri, 19 Nov 2021 14:35:27 +0000 (22:35 +0800)]
oeqa: fix warnings for append operators combined with +=

Signed-off-by: Yi Zhao <yi.zhao@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agorust-cross: Replace TARGET_ARCH with TUNE_PKGARCH
Pgowda [Mon, 15 Nov 2021 14:03:40 +0000 (06:03 -0800)]
rust-cross: Replace TARGET_ARCH with TUNE_PKGARCH

rust-cross-* imported from meta-rust has incorrect signatures,
depending on MACHINEOVERRIDES making it effectively MACHINE_ARCH
as shown by sstate-diff-machines.sh:

openembedded-core/scripts/sstate-diff-machines.sh --tmpdir=tmp-glibc \
--machines="qemuarm64 qemuarm64copy" --targets=rust-cross-aarch64-glibc \
--analyze

 === Comparing signatures for task do_configure.sigdata between
 qemuarm64 and qemuarm64copy ===
ERROR: gcc-runtime different signature for task do_configure.sigdata
between qemuarm64 and qemuarm64copy
NOTE: Starting bitbake server...
Hash for dependent task gcc/gcc-runtime_11.2.bb:do_prepare_recipe_sysroot
changed from
da4ebf1b272cb73153145a0a95e6438d2955ae2d36f84db10f6880b2781ec331 to
47a0ebb7a88c9f896fb9dbce269f575ab8a6faabb2b9e62d164be6e71c5e4e40
Unable to find matching sigdata for
openembedded-core/meta/recipes-devtools/gcc/gcc-runtime_11.2.bb:
do_prepare_recipe_sysroot with hashes
da4ebf1b272cb73153145a0a95e6438d2955ae2d36f84db10f6880b2781ec331 or
47a0ebb7a88c9f896fb9dbce269f575ab8a6faabb2b9e62d164be6e71c5e4e40

The following patch takes TUNE_PKGARCH into consideration instead
of TARGET_ARCH and results in signatures as expected.

[YOCTO #14613]

RP: Added maintainer.inc corresponding change
Signed-off-by: Pgowda <pgowda.cve@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agowic: use shutil.which
Mingli Yu [Wed, 17 Nov 2021 02:16:23 +0000 (10:16 +0800)]
wic: use shutil.which

Use shutil.which to find the executable instead to silence the below warning:
 $ cat tmp/work/intel_x86_64-poky-linux/core-image-base/1.0-r5/temp/log.do_image_wic
 [snip]
 DEBUG: Executing shell function do_image_wic
 /path/layers/oe-core/scripts/wic:27: DeprecationWarning: The distutils package is deprecated and slated for removal in Python 3.12. Use setuptools or check PEP 632 for potential alternatives
  from distutils import spawn
 INFO: Creating image(s)...
 [snip]

[RP: Added conversion for missed function reference]
Signed-off-by: Mingli Yu <mingli.yu@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agogstreamer1.0-plugins-base: add support for graphene
Ulrich Ölmann [Thu, 18 Nov 2021 21:33:00 +0000 (22:33 +0100)]
gstreamer1.0-plugins-base: add support for graphene

Signed-off-by: Ulrich Ölmann <u.oelmann@pengutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agocreate-spdx: Fix key errors in do_create_runtime_spdx
Andres Beltran [Thu, 18 Nov 2021 00:04:02 +0000 (00:04 +0000)]
create-spdx: Fix key errors in do_create_runtime_spdx

Currently, the do_create_runtime_spdx task fails with a Key Error if a
dependency is not contained in the package providers dictionary. Add a
check before using "dep" as a key in "providers".

Signed-off-by: Andres Beltran <abeltran@linux.microsoft.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agocreate-spdx: Protect against None from LICENSE_PATH
Saul Wold [Wed, 17 Nov 2021 18:08:14 +0000 (10:08 -0800)]
create-spdx: Protect against None from LICENSE_PATH

If LICENSE_PATH is not set, then the split() will fail on a NoneType.

Signed-off-by: Saul Wold <saul.wold@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agocreate_spdx: ensure is_work_shared() is unique
Saul Wold [Wed, 17 Nov 2021 18:08:13 +0000 (10:08 -0800)]
create_spdx: ensure is_work_shared() is unique

There is a function with the same name is_work_shared() in the archiver class
this causes a conflict when both classes are included. Use work-shared as the
check in WORKDIR to allow for other packages beyond the kernel and gcc that
use a common shared-work source directory.

Signed-off-by: Saul Wold <saul.wold@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agoos-release: Add DISTRO_CODENAME as vardeps for do_compile
Daniel Gomez [Thu, 18 Nov 2021 22:06:09 +0000 (23:06 +0100)]
os-release: Add DISTRO_CODENAME as vardeps for do_compile

DISTRO_CODENAME is part of VERSION variable but not used as dependency
for do_compile task. Append it to the vardeps list to rebuild in case it
changes.

Signed-off-by: Daniel Gomez <daniel@qtec.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agolibrsvg: submit system-deps patch upstream
Alexander Kanavin [Wed, 17 Nov 2021 15:35:25 +0000 (16:35 +0100)]
librsvg: submit system-deps patch upstream

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agobusybox: drop 0001-Use-CC-when-linking-instead-of-LD-and-use-CFLAGS-and.patch
Alexander Kanavin [Wed, 17 Nov 2021 15:35:23 +0000 (16:35 +0100)]
busybox: drop 0001-Use-CC-when-linking-instead-of-LD-and-use-CFLAGS-and.patch

This was added 10 years ago, is almost certainly non-upstreamable
and it isn't clear what the issues it aims to fix are:
the AB revealed no problems when the patch is removed.

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agoslang: mark patch as inappropriate for upstream submission
Alexander Kanavin [Wed, 17 Nov 2021 15:35:24 +0000 (16:35 +0100)]
slang: mark patch as inappropriate for upstream submission

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agoexpect: do not install examples, drop related patch
Alexander Kanavin [Wed, 17 Nov 2021 15:35:22 +0000 (16:35 +0100)]
expect: do not install examples, drop related patch

There is really no reason to install them; this was done
in the original recipe submission for unexplained reasons.

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agonumactl: add missing Upstream-Status to a patch
Alexander Kanavin [Wed, 17 Nov 2021 15:35:20 +0000 (16:35 +0100)]
numactl: add missing Upstream-Status to a patch

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agorust-hello-world: add missing Upstream-Status to a patch
Alexander Kanavin [Wed, 17 Nov 2021 15:35:19 +0000 (16:35 +0100)]
rust-hello-world: add missing Upstream-Status to a patch

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agolibnl: add missing Upstream-Status tag
Alexander Kanavin [Wed, 17 Nov 2021 15:35:17 +0000 (16:35 +0100)]
libnl: add missing Upstream-Status tag

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agokea: add missing Upstream-Status to a patch
Alexander Kanavin [Wed, 17 Nov 2021 15:35:18 +0000 (16:35 +0100)]
kea: add missing Upstream-Status to a patch

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agounfs3: add missing Upstream-Status tag
Alexander Kanavin [Wed, 17 Nov 2021 15:35:16 +0000 (16:35 +0100)]
unfs3: add missing Upstream-Status tag

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agolibxml2: add missing Upstream-Status tag
Alexander Kanavin [Wed, 17 Nov 2021 15:35:15 +0000 (16:35 +0100)]
libxml2: add missing Upstream-Status tag

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agoautoconf: add missing Upstream-Status to patches
Alexander Kanavin [Wed, 17 Nov 2021 15:35:14 +0000 (16:35 +0100)]
autoconf: add missing Upstream-Status to patches

All of them seem like very old, very oe specific tweaks; please adjust otherwise.

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agopatches: correct whitespace/spelling for Upstream-Status tags
Alexander Kanavin [Wed, 17 Nov 2021 15:35:13 +0000 (16:35 +0100)]
patches: correct whitespace/spelling for Upstream-Status tags

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agolibsamplerate0: update 0.1.9 -> 0.2.2
Alexander Kanavin [Wed, 17 Nov 2021 15:35:12 +0000 (16:35 +0100)]
libsamplerate0: update 0.1.9 -> 0.2.2

Upstream moved to github.

Drop both patches: alsa configuration merged upstream,
shared version info fixed upstream.

License-Update: urls changed to github.

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agolibrsvg: update 2.52.3 -> 2.52.4
Alexander Kanavin [Wed, 17 Nov 2021 15:35:09 +0000 (16:35 +0100)]
librsvg: update 2.52.3 -> 2.52.4

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agocurl: update 7.79.1 -> 7.80.0
Alexander Kanavin [Wed, 17 Nov 2021 15:35:08 +0000 (16:35 +0100)]
curl: update 7.79.1 -> 7.80.0

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agolibrstd-rs: submit patches upstream
Alexander Kanavin [Wed, 17 Nov 2021 15:35:10 +0000 (16:35 +0100)]
librstd-rs: submit patches upstream

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agoxwayland: upgrade 21.1.2 -> 21.1.3
Alexander Kanavin [Wed, 17 Nov 2021 15:35:07 +0000 (16:35 +0100)]
xwayland: upgrade 21.1.2 -> 21.1.3

Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agosystemd: upgrade 249.5 -> 249.6
Alexander Kanavin [Wed, 17 Nov 2021 15:35:06 +0000 (16:35 +0100)]
systemd: upgrade 249.5 -> 249.6

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agopiglit: upgrade to latest revision
Alexander Kanavin [Wed, 17 Nov 2021 15:35:04 +0000 (16:35 +0100)]
piglit: upgrade to latest revision

Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agodiffoscope: upgrade 190 -> 192
Alexander Kanavin [Wed, 17 Nov 2021 15:35:05 +0000 (16:35 +0100)]
diffoscope: upgrade 190 -> 192

Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agolibpipeline: upgrade 1.5.3 -> 1.5.4
Alexander Kanavin [Wed, 17 Nov 2021 15:35:03 +0000 (16:35 +0100)]
libpipeline: upgrade 1.5.3 -> 1.5.4

License-Update: http -> https

Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agoasciidoc: upgrade 10.0.1 -> 10.0.2
Alexander Kanavin [Wed, 17 Nov 2021 15:35:02 +0000 (16:35 +0100)]
asciidoc: upgrade 10.0.1 -> 10.0.2

Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agopython3-pyparsing: upgrade 3.0.5 -> 3.0.6
Alexander Kanavin [Wed, 17 Nov 2021 15:35:01 +0000 (16:35 +0100)]
python3-pyparsing: upgrade 3.0.5 -> 3.0.6

Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agopython3-hypothesis: upgrade 6.24.2 -> 6.24.3
Alexander Kanavin [Wed, 17 Nov 2021 15:35:00 +0000 (16:35 +0100)]
python3-hypothesis: upgrade 6.24.2 -> 6.24.3

Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agoccache: upgrade 4.4.2 -> 4.5
Alexander Kanavin [Wed, 17 Nov 2021 15:34:59 +0000 (16:34 +0100)]
ccache: upgrade 4.4.2 -> 4.5

License-Update: third party version references updated

Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agogettext: address automake 1.16.5 failures
Alexander Kanavin [Wed, 17 Nov 2021 15:34:58 +0000 (16:34 +0100)]
gettext: address automake 1.16.5 failures

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agolibunistring: address automake 1.16.5 failures
Alexander Kanavin [Wed, 17 Nov 2021 15:34:57 +0000 (16:34 +0100)]
libunistring: address automake 1.16.5 failures

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agoautomake: update 1.16.3 -> 1.16.5
Alexander Kanavin [Wed, 17 Nov 2021 15:34:56 +0000 (16:34 +0100)]
automake: update 1.16.3 -> 1.16.5

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agounzip/zip: mark all patches as non-upstreamable
Alexander Kanavin [Wed, 17 Nov 2021 15:34:55 +0000 (16:34 +0100)]
unzip/zip: mark all patches as non-upstreamable

While upstream is not technically dead as of today [1], it has not made
a release in over a decade, and there is no source code repo to
rebase the patches on. If/when a new release ever comes, or someone
takes over the maintenance, we can reassess the situation with these patches.

[1] Discussion in https://sourceforge.net/p/infozip/bugs/53/

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agodb: mark all patches as non-upstreamable
Alexander Kanavin [Wed, 17 Nov 2021 15:34:54 +0000 (16:34 +0100)]
db: mark all patches as non-upstreamable

Since Oracle relicensed bdb, the open source community
is slowly but surely replacing bdb with supported and
open source friendly alternatives.

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agoweston: submit patch upstream
Alexander Kanavin [Wed, 17 Nov 2021 15:34:53 +0000 (16:34 +0100)]
weston: submit patch upstream

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agoovmf: submit patch upstream
Alexander Kanavin [Wed, 17 Nov 2021 15:34:50 +0000 (16:34 +0100)]
ovmf: submit patch upstream

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agoscripts/lnr: remove
Ross Burton [Wed, 17 Nov 2021 17:23:08 +0000 (17:23 +0000)]
scripts/lnr: remove

lnr is a script that implements the same behaviour as 'ln --relative
--symlink', as at the time of creation --relative was only available in
coreutils 8.16 onwards which was too new for the older supported distros.

Now, however, everyone has a new enough coreutils, so we can remove this
script.

All users of lnr should be replaced with ln --relative --symbolic.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agogdb-common: Add --with-liblzma-prefix in xz PACKAGECONFIG.
Jim Wilson [Tue, 16 Nov 2021 01:44:13 +0000 (17:44 -0800)]
gdb-common: Add --with-liblzma-prefix in xz PACKAGECONFIG.

This fixes an arch linux gdb configure error with liblzma installed.

Without the --with-liblzma-prefix option, when configuring the target
gdb, gdb configure may find the native liblzma (on a non-multiarch
system) and try to use it, which gives a configure error.  We already
use --with-libexpat-prefix for expat, we just need to do the same for
liblzma.

This was reported here:
https://github.com/sifive/meta-sifive/issues/34
https://github.com/sifive/freedom-u-sdk/issues/148

Signed-off-by: Jim Wilson <jim.wilson.gcc@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agoethtool: upgrade 5.14 -> 5.15
wangmy [Tue, 16 Nov 2021 01:28:41 +0000 (09:28 +0800)]
ethtool: upgrade 5.14 -> 5.15

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agogcc: Tweak Upstream-Status formatting
Richard Purdie [Fri, 19 Nov 2021 11:04:04 +0000 (11:04 +0000)]
gcc: Tweak Upstream-Status formatting

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agolibtasn1: upgrade 4.17.0 -> 4.18.0
wangmy [Mon, 15 Nov 2021 15:39:10 +0000 (23:39 +0800)]
libtasn1: upgrade 4.17.0 -> 4.18.0

* Noteworthy changes in release 4.18.0 (2021-11-09) [stable]
- Improve GTK-DOC manual.  Closes: #35.
- Improve --help and --version for tools with gnulib.  Closes: #37.
- Update gnulib files and various maintenance fixes.

refresh dont-depend-on-help2man.patch

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agovim: fix CVE-2021-3927 and CVE-2021-3928
Ross Burton [Mon, 15 Nov 2021 12:53:28 +0000 (12:53 +0000)]
vim: fix CVE-2021-3927 and CVE-2021-3928

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agopython3-more-itertools: upgrade 8.10.0 -> 8.11.0
wangmy [Mon, 15 Nov 2021 15:04:14 +0000 (23:04 +0800)]
python3-more-itertools: upgrade 8.10.0 -> 8.11.0

8.11.0:
New functions
The :func:`before_and_after`, :func:`sliding_window`, and :func:`triplewise` recipes from the Python 3.10 docs were added
:func:`duplicates_everseen` and :func:`duplicates_justseen` (thanks to OrBin and DavidPratt512)
:func:`minmax` (thanks to Ricocotam, MSeifert04, and ruancomelli)
:func:`strictly_n` (thanks to hwalinga and NotWearingPants)
:func:`unique_in_window`
Changes to existing functions
:func:`groupby_transform` had its type stub improved (thanks to mjk4 and ruancomelli)
:func:`is_sorted` now accepts a strict parameter (thanks to Dutcho and ruancomelli)
:func:`zip_broadcast` was updated to fix a bug (thanks to kalekundert)

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agopython3-jinja2: upgrade 3.0.2 -> 3.0.3
wangmy [Mon, 15 Nov 2021 15:04:13 +0000 (23:04 +0800)]
python3-jinja2: upgrade 3.0.2 -> 3.0.3

What's Changed
update requirements by @davidism in #1524
test Python 3.10 by @davidism in #1530
test Python 3.11 by @davidism in #1531
Tests: Adapt expected traceback regexes for Python 3.11.0a1 by @hroncok in #1527
simplify CodeType rewriting by @davidism in #1536
native: keep same behavior on Python 3.10 by @mkrizek in #1537
Reduce async overhead due to auto_await by @mohd-akram in #1516
Use object.__hash__ for Node.__hash__ by @klette in #1522
Fix typos by @kianmeng in #1509
rewrite Template class doc by @davidism in #1538
remove reference to macro.defaults by @davidism in #1539
rewrite docs about extending template objects by @davidism in #1540
document chained pow order by @davidism in #1541
PackageLoader works with single module file by @davidism in #1542
release version 3.0.3 by @davidism in #1543

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agopython3-importlib-metadata: upgrade 4.8.1 -> 4.8.2
wangmy [Mon, 15 Nov 2021 15:03:34 +0000 (23:03 +0800)]
python3-importlib-metadata: upgrade 4.8.1 -> 4.8.2

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agolibhandy: upgrade 1.4.0 -> 1.5.0
wangmy [Mon, 15 Nov 2021 15:02:36 +0000 (23:02 +0800)]
libhandy: upgrade 1.4.0 -> 1.5.0

=============
Version 1.5.0
=============

- Introduce HdyStyleManager for managing color schemes (light/dark) and
  high contrast mode.
  - Support the cross-platform color scheme preference in the
    settings portal.
- HdyPreferencesPage
  - Remove adaptive margins to fix default window sizes
- HdyTabBar
  - Handle middle click on button release rather than press
- Various memory leak fixes
- Translation updates:
  - Brazilian Portuguese
  - Finnish
  - Indonesian
  - Portuguese
  - Serbian
  - Swedish
  - Ukrainian

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agolibdrm: upgrade 2.4.107 -> 2.4.108
wangmy [Mon, 15 Nov 2021 15:02:34 +0000 (23:02 +0800)]
libdrm: upgrade 2.4.107 -> 2.4.108

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agolibevdev: upgrade 1.11.0 -> 1.12.0
wangmy [Mon, 15 Nov 2021 15:02:35 +0000 (23:02 +0800)]
libevdev: upgrade 1.11.0 -> 1.12.0

determinism.patch removed since it is included in 1.12.0

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agospirv-headers: PV should be 1.5.4
Alexander Kanavin [Mon, 15 Nov 2021 14:06:30 +0000 (15:06 +0100)]
spirv-headers: PV should be 1.5.4

As upstream hasn't actually tagged 1.5.5, we should
continue to use 1.5.4 until it does.

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agoglibc: Backport fix for CVE-2021-43396
Richard Purdie [Mon, 15 Nov 2021 11:49:06 +0000 (11:49 +0000)]
glibc: Backport fix for CVE-2021-43396

Backport the fix for CVE-2021-43396. It is disputed that this is a security issue
however the fix applies easily so we may as well.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agolua: Backport fix for CVE-2021-43396
Richard Purdie [Mon, 15 Nov 2021 11:47:35 +0000 (11:47 +0000)]
lua: Backport fix for CVE-2021-43396

Backport the fix for CVE-2021-43396 ("C stack overflow with coroutines")
from upstream.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agowaffle: add cmake-native to DEPENDS
Chen Qi [Mon, 15 Nov 2021 09:04:40 +0000 (01:04 -0800)]
waffle: add cmake-native to DEPENDS

waffle currently requires cmake-native because it has the following lines
in git/src/waffle/meson.build.

  if meson.version().version_compare('>= 0.50')
    cmake = import('cmake')

In case of 'wayland' not being in DISTRO_FEATURES, we will get the
following error at do_configure.

  | ../git/src/waffle/meson.build:233:8: ERROR: Unable to find cmake

So explictly add 'cmake-native' to DEPENDS to solve this problem.

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agogcc: Dropping mips workaround
Richard Purdie [Thu, 11 Nov 2021 17:00:50 +0000 (17:00 +0000)]
gcc: Dropping mips workaround

I've tested without this and the ptest results for mips are the same with
and without it so the issue this was fixing in gcc 9 was likely resolved
by gcc 11.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agogcc: Fix CVE-2021-35465
Pgowda [Mon, 15 Nov 2021 11:18:21 +0000 (03:18 -0800)]
gcc: Fix CVE-2021-35465

source : https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102035

Upstream-Status: Backport[https://gcc.gnu.org/git/gitweb.cgi?p=gcc.git;h=3929bca9ca95de9d35e82ae8828b188029e3eb70]
Upstream-Status: Backport[https://gcc.gnu.org/git/gitweb.cgi?p=gcc.git;h=574e7950bd6b34e9e2cacce18c802b45505d1d0a]
Upstream-Status: Backport[https://gcc.gnu.org/git/gitweb.cgi?p=gcc.git;h=30461cf8dba3d3adb15a125e4da48800eb2b9b8f]
Upstream-Status: Backport[https://gcc.gnu.org/git/gitweb.cgi?p=gcc.git;h=809330ab8450261e05919b472783bf15e4b000f7]

Signed-off-by: Pgowda <pgowda.cve@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agosstate: add a function to get the sstate file name in sstate_checkhashes
Jose Quaresma [Tue, 9 Nov 2021 11:11:48 +0000 (11:11 +0000)]
sstate: add a function to get the sstate file name in sstate_checkhashes

Signed-off-by: Jose Quaresma <quaresma.jose@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agosstate: only search on the mirrors for the missing files
Jose Quaresma [Tue, 9 Nov 2021 11:11:47 +0000 (11:11 +0000)]
sstate: only search on the mirrors for the missing files

The existence of missed files on the local sstate cache is the
condition to search for it on the sstate mirror.
The missed collection are populated when we search the files
on the local sstate cache.

- When we have sstate mirrors we don't need to iterate all files
again because we already know what's missing.

- When we found a file on the sstate mirror is because of they
are missing on the local sstate mirror.
So avoid checking if the hash exists on the missed collection.

- When we can't find the file on the sstate mirror we don't need
to add the hash again to the missed collection as it already exists.

Signed-off-by: Jose Quaresma <quaresma.jose@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agosstate: cleanup the sstate mirror progress check
Jose Quaresma [Tue, 9 Nov 2021 11:11:46 +0000 (11:11 +0000)]
sstate: cleanup the sstate mirror progress check

We only have the progress bar when there are more than 100 objects.
So we can check for this and store the result to show the progress bar.

Signed-off-by: Jose Quaresma <quaresma.jose@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agoshaderc: upgrade 2021.2 -> 2021.3
Jose Quaresma [Sun, 14 Nov 2021 20:26:48 +0000 (20:26 +0000)]
shaderc: upgrade 2021.2 -> 2021.3

- SHADERC_SKIP_COPYRIGHT_CHECK is a new cmake option to skip
the copyright check and as we don't needs it we can avoid the check.
https://github.com/google/shaderc/commit/f6d6dddfabfec1041c0dfb8e7ff3608a5f82227c

Signed-off-by: Jose Quaresma <quaresma.jose@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agoglslang: upgrade 11.6.0 -> 11.7.0
Jose Quaresma [Sun, 14 Nov 2021 20:26:47 +0000 (20:26 +0000)]
glslang: upgrade 11.6.0 -> 11.7.0

Signed-off-by: Jose Quaresma <quaresma.jose@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agospirv-tools: upgrade 2021.3 -> 2021.4
Jose Quaresma [Sun, 14 Nov 2021 20:26:46 +0000 (20:26 +0000)]
spirv-tools: upgrade 2021.3 -> 2021.4

Signed-off-by: Jose Quaresma <quaresma.jose@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agospirv-headers: bump to last master
Jose Quaresma [Sun, 14 Nov 2021 20:26:45 +0000 (20:26 +0000)]
spirv-headers: bump to last master

SPIRV-Tools v2021.4 require this revision
https://github.com/KhronosGroup/SPIRV-Tools/commit/21e3f681e2004590c7865bc8c0195a4ab8e66c88

Signed-off-by: Jose Quaresma <quaresma.jose@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agospirv-headers: upgrade 1.5.4 -> 1.5.5
Jose Quaresma [Sun, 14 Nov 2021 20:26:44 +0000 (20:26 +0000)]
spirv-headers: upgrade 1.5.4 -> 1.5.5

Signed-off-by: Jose Quaresma <quaresma.jose@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agovulkan-samples: Enable for architectures where it builds
Khem Raj [Sun, 14 Nov 2021 16:52:35 +0000 (08:52 -0800)]
vulkan-samples: Enable for architectures where it builds

With last upgade it has stopped working on arm/mips/x86/rv32
basically all 32bit architectures. Until this is fixed enable it
for architectures where its known to build now.

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agovulkan-samples: Fix cmake errors about plain signature
Khem Raj [Sun, 14 Nov 2021 16:52:34 +0000 (08:52 -0800)]
vulkan-samples: Fix cmake errors about plain signature

Fix some typecasting errors seen on 32bit platforms with clang

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agopython3-numpy: add python3-json to RDEPENDS
Yi Zhao [Sun, 14 Nov 2021 05:15:56 +0000 (13:15 +0800)]
python3-numpy: add python3-json to RDEPENDS

Fixes:
$ python3
>>> import numpy
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/lib/python3.10/site-packages/numpy/__init__.py", line 138, in <module>
    from ._version import get_versions
  File "/usr/lib/python3.10/site-packages/numpy/_version.py", line 7, in <module>
    import json
ModuleNotFoundError: No module named 'json'
>>>

Signed-off-by: Yi Zhao <yi.zhao@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agopython3-numpy: move recipe to python directory
Yi Zhao [Sun, 14 Nov 2021 05:15:55 +0000 (13:15 +0800)]
python3-numpy: move recipe to python directory

This recipe had been moved out from python directory since 2016[1] in
order to share patches between python2 and python3. But now there is no
reason to keep it in its own directory as we only keep python3-nump.
Move it back to python directory.

[1] https://git.openembedded.org/openembedded-core/commit/?id=9bffe2f9fb4ce6c0b265f27e5b484fbe076c6349

Signed-off-by: Yi Zhao <yi.zhao@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agocups: Fix missing installation of cups sysv init scripts
Claus Stovgaard [Sat, 13 Nov 2021 21:37:13 +0000 (22:37 +0100)]
cups: Fix missing installation of cups sysv init scripts

The packageconfig needs to be --disable-systemd as documented in
configure file for cups. With the current value "--without-systemd" the
SYSTEM_DIR variable ends up being set to "no"

It is caused by the --without-* section in configure file resulting in
eval with_$ac_useropt=no ;;

$ac_useropt is "systemd" causing the variable $with_systemd to be set
to "no", because of below test

if test ${with_systemd+y}
then :
  withval=$with_systemd; SYSTEMD_DIR="$withval"
else $as_nop
  SYSTEMD_DIR=""
fi

cups configure test for i if SYSTEMD_DIR is empty to decide if the init
scripts need to be installed. A value of "no" results in that no init
scripts is installed.

With --disable-systemd it works as expected - installing the init files.
Though cups should properly improve their configure script.

Signed-off-by: Claus Stovgaard <clst@ambu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agopackage: Add sanity check for allarch packagegroups
Richard Purdie [Fri, 12 Nov 2021 16:33:14 +0000 (16:33 +0000)]
package: Add sanity check for allarch packagegroups

We exclude allarch packagegroups from rebuilding when their dependencies change.
The reasoning is that we are just depending on a name so having these rebuild
lots is just pointless and inefficient. We also don't want them duplicated for
multiple machines for efficiency.

In general this works fine, as long as the package names don't change. That
is also rare but there is one corner case which does catch users out - debian
package renaming. When this does break, users question sstate and so on and
lose faith in the system even if this is a known choice we made.

This commit adds an error message if an allarch packagegroup depends on any
package which shows package renaming in action (through the PKG variable
being set).

If you run into this issue you either need to remove the dependency from the
packagegroup or mark the packagegroup as tune specific, i.e. set:

PACKAGE_ARCH = "${TUNE_PKGARCH}"

before the packagegroup inherit.

[YOCTO #7298]

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agorpcbind: install rpcbind.conf
Yi Zhao [Sat, 13 Nov 2021 07:47:01 +0000 (15:47 +0800)]
rpcbind: install rpcbind.conf

Although we provide rpcbind.conf, it isn't shipped to the package.
Also fix the sysconfdir in rpcbind.service.

Signed-off-by: Yi Zhao <yi.zhao@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agorunqemu: support rootfs mounted ro
Adrian Freihofer [Fri, 12 Nov 2021 22:29:18 +0000 (23:29 +0100)]
runqemu: support rootfs mounted ro

Optionally allow to set QB_KERNEL_ROOT to e.g. "/dev/vda ro" to mount the
rootfs reay-only in Qemu.

Signed-off-by: Adrian Freihofer <adrian.freihofer@siemens.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agopython3-pygobject: Improve introspection handling
Khem Raj [Fri, 12 Nov 2021 21:03:45 +0000 (13:03 -0800)]
python3-pygobject: Improve introspection handling

meson 0.60 turns unknown options into them hard errors. Do not add
-Dintrospection by using the new configuration option.

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agogstreamer1.0-python: Improve introspection option handling
Khem Raj [Fri, 12 Nov 2021 21:03:44 +0000 (13:03 -0800)]
gstreamer1.0-python: Improve introspection option handling

meson 0.60 turns unknown options into them hard errors. Disable
the addition of -Dintrospection using the new configuration option.

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agomirrors: Add kernel.org sources mirror for downloads.yoctoproject.org
Richard Purdie [Fri, 12 Nov 2021 17:48:25 +0000 (17:48 +0000)]
mirrors: Add kernel.org sources mirror for downloads.yoctoproject.org

kernel.org now has a mirror of the downloads.yoctoproject.org sources
archive so include this in our mirrors list.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agogobject-introspection: Add -Dintrospection only when GIR_MESON_OPTION is not empty
Khem Raj [Fri, 12 Nov 2021 00:59:23 +0000 (16:59 -0800)]
gobject-introspection: Add -Dintrospection only when GIR_MESON_OPTION is not empty

meson now does not allow unknown options which means we have to be
mindful of adding -Dintrospection unconditionally and provide an option
to recipes to set GIR_MESON_OPTION = "" and when recipe does that then
do not add -D<Dintrospection> option to meson. This will help recipes to
use meson 60.x+ just by adding

GIR_MESON_OPTION = ""

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Cc: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agopackagegroup-core-x11-base: Drop dbus dependency
Richard Purdie [Fri, 12 Nov 2021 16:40:40 +0000 (16:40 +0000)]
packagegroup-core-x11-base: Drop dbus dependency

dbus isn't an x11 dependency and this suffers from debian renaming. Simply drop
the dependency.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agopackagegroup-core-full-cmdline: Drop compatibility mappings
Richard Purdie [Fri, 12 Nov 2021 17:12:14 +0000 (17:12 +0000)]
packagegroup-core-full-cmdline: Drop compatibility mappings

The task-core namespace was dropped years ago and we've had the compatibility
mappings for a long time. We should no longer need them as everyone should long
since have adapted.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 years agopackagegroups-core-full-cmdline: Drop libraries packagegroup and gmp
Richard Purdie [Fri, 12 Nov 2021 17:01:27 +0000 (17:01 +0000)]
packagegroups-core-full-cmdline: Drop libraries packagegroup and gmp

We pull in libraries as/where needed as dependencies so there is no need
to have them as specific packagegroups. This change removes glib-2.0 and gmp.

This also has the advantage of meaning debian renaming now isn't used anywhere
and the packagegroup can remain allarch.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>