]> git.ipfire.org Git - thirdparty/openembedded/openembedded-core.git/log
thirdparty/openembedded/openembedded-core.git
2 years agolttng-modules: fix for v6.3+ kernels
Bruce Ashfield [Sat, 18 Mar 2023 15:32:02 +0000 (11:32 -0400)] 
lttng-modules: fix for v6.3+ kernels

As part of updating linux-yocto-dev to v6.3, lttng-modules needs to
have four backported patches to adjust for the new kernel version.

These are safe for inclusion in the upcoming release, as they are
version protected and provide some future proofing for folks trying
to use newer kernels than our released 6.1.

Signed-off-by: Bruce Ashfield <bruce.ashfield@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agolttng-modules: update to v2.13.9
Bruce Ashfield [Sat, 18 Mar 2023 15:32:01 +0000 (11:32 -0400)] 
lttng-modules: update to v2.13.9

Bumping lttng-modules to version v2.13.9-4-g12f43cab, which comprises the following commits:

    da1f5a26 Version 2.13.9
    dc2d1294 fix: jbd2: use the correct print format (v5.4.229)
    d04c1211 fix: jbd2 upper bound for v5.10.163
    4b8864fc fix: jbd2: use the correct print format (v5.10.163)
    69d3aa79 fix: btrfs: move accessor helpers into accessors.h (v6.2)

We drop our previously backported commits as well, since they are part
of the release.

Signed-off-by: Bruce Ashfield <bruce.ashfield@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agomdadm: fix tests/00raid0
Mingli Yu [Fri, 17 Mar 2023 12:37:42 +0000 (20:37 +0800)] 
mdadm: fix tests/00raid0

Backport patch [1] to fix tests/00raid0.

[1] https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/commit/?id=7539254342bc591717b0051734cc6c09c1b88640

Signed-off-by: Mingli Yu <mingli.yu@windriver.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agoglibc: Disable warnings as errors
Khem Raj [Fri, 17 Mar 2023 13:18:32 +0000 (06:18 -0700)] 
glibc: Disable warnings as errors

As distro builders this option helps using different compilers to build
glibc

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agomesa: update 22.3.5 -> 23.0.0
Alexander Kanavin [Thu, 16 Mar 2023 09:41:02 +0000 (10:41 +0100)] 
mesa: update 22.3.5 -> 23.0.0

Drop patches:
0001-gbm-backend-fix-gbm-compile-without-dri.patch
0001-freedreno-pm4-Use-unsigned-instead-of-uint-to-fix-mu.patch
(backports)

0001-util-format-Check-for-NEON-before-using-it.patch
(upstream submission has stalled with unanswered concerns from upstream;
patch no longer applies; issue does not reproduce with 'raspberrypi0-wifi' MACHINE)

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
2 years agomesa: do not strip rpaths from dri drivers
Alexander Kanavin [Thu, 16 Mar 2023 09:41:01 +0000 (10:41 +0100)] 
mesa: do not strip rpaths from dri drivers

This was needed back when we used upstream llvm-config which
would add bogus rpaths into the drivers; nowadays it is replaced
with a oe-specific wrapper.

The particular reason is that mesa-native relies on having rpaths
to work correctly (i.e. find other native libraries in the
same sysroot location).

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
2 years agollvm: allow building libllvm in native builds, subject to PACKAGECONFIG
Alexander Kanavin [Thu, 16 Mar 2023 09:41:00 +0000 (10:41 +0100)] 
llvm: allow building libllvm in native builds, subject to PACKAGECONFIG

Also, enable that, if opengl is in native DISTRO_FEATURES: this allows
mesa-native to build drivers that rely on libllvm, particularly llvmpipe,
which is a (sort of) accelerated software renderer that is the only
option on build hosts without dedicated or supported GPUs.

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
2 years agomesa: enable a rich set of drivers for native builds
Alexander Kanavin [Thu, 16 Mar 2023 09:40:59 +0000 (10:40 +0100)] 
mesa: enable a rich set of drivers for native builds

Without this, the only available driver would be swrast
(without llvmpipe) which is far too slow to be practical,
and so qemu wouldn't be able to perform accelerated graphics.

The one major dependency pulled in by this change is llvm-native
which takes just under three minutes on my machine. If this is
undesitable, opengl can be removed from DISTRO_FEATURES_NATIVE.

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
2 years agomesa: allow mesa-native/nativesdk only subject to opengl/vulkan DISTRO_FEATURE
Alexander Kanavin [Thu, 16 Mar 2023 09:40:58 +0000 (10:40 +0100)] 
mesa: allow mesa-native/nativesdk only subject to opengl/vulkan DISTRO_FEATURE

This prevents accidental builds of it when native/nativesdk opengl
is not actually enabled (which may cause undesirable lengthening
of dependency chains or build failures because those dependencies
don't enable options required for mesa).

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
2 years agorunqemu: direct mesa to use its own drivers, rather than ones provided by host distro
Alexander Kanavin [Thu, 16 Mar 2023 09:40:57 +0000 (10:40 +0100)] 
runqemu: direct mesa to use its own drivers, rather than ones provided by host distro

With mesa 23.0, it is not longer possible to use the host drivers, as
mesa upstream has added strict checks for matching builds between
drivers and libraries that load them.

Add a check and a hint to runqemu so that there is a helpful error when
there is no native/nativesdk opengl/virgl support.

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
2 years agooeqa: loader.py: show warning when skipping selected module and abort if all are...
Martin Jansa [Mon, 13 Mar 2023 14:20:31 +0000 (15:20 +0100)] 
oeqa: loader.py: show warning when skipping selected module and abort if all are skipped

* skipped modules were triggering an ERROR before:

  poky $ oe-selftest -v -r eSDK.oeSDKExtSelfTest.test_install_libraries_headers imagefeatures.ImageFeatures.test_image_gen_debugfs -K -B /OE/build/poky/build-eSDK
  2023-03-13 15:07:53,430 - oe-selftest - ERROR - Not found eSDK.oeSDKExtSelfTest.test_install_libraries_headers in loaded test cases

* but didn't show the reason why it wasn't loaded and more importantly -r
  was ignored when all selected modules were silently skipped

* add a warning when skipping some module and abort if some modules were
  selected, but all ended being skipped:

  poky $ oe-selftest -v -r eSDK.oeSDKExtSelfTest.test_install_libraries_headers -K -B /OE/build/poky/build-eSDK
  2023-03-13 15:11:51,028 - oe-selftest - WARNING - module 'eSDK.oeSDKExtSelfTest.test_install_libraries_headers' was skipped from selected modules, because it doesn't match with module name assumptions: package and module names do not contain upper case characters, whereas class names do
  2023-03-13 15:11:51,028 - oe-selftest - ERROR - All selected modules were skipped, this would trigger selftest with all tests and -r ignored.

* I was hit by this in oe-selftest -r eSDK.oeSDKExtSelfTest.test_install_libraries_headers
  which is skipped due to upper case characters in module name and selftest started to run
  all tests (archiver.Archiver.test_archiver_allows_to_filter_on_recipe_name as first from 529)

  poky $ oe-selftest -v -r eSDK.oeSDKExtSelfTest.test_install_libraries_headers -K -B /OE/build/poky/build-eSDK
  2023-03-13 14:00:52,955 - oe-selftest - DEBUG - Selected tests with -r: ['eSDK.oeSDKExtSelfTest.test_install_libraries_headers']
  2023-03-13 14:00:55,531 - oe-selftest - INFO - Changing cwd to /OE/build/poky/build
  ..
  2023-03-13 14:00:58,128 - oe-selftest - INFO - test_archiver_allows_to_filter_on_recipe_name (archiver.Archiver.test_archiver_allows_to_filter_on_recipe_name)

  I'll rename eSDK to esdk in next commit to avoid this.

* also fix small typo in context I've noticed when debugging this

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
2 years agoselftest: imagefeatures.py: respect IMAGE_LINK_NAME for debugfs and manifest as well
Martin Jansa [Mon, 13 Mar 2023 12:15:41 +0000 (13:15 +0100)] 
selftest: imagefeatures.py: respect IMAGE_LINK_NAME for debugfs and manifest as well

* these cases were correctly respecting IMAGE_LINK_NAME in most tests

* the only exception was relatively wide glob for manifest:
  "test-empty-image-*.manifest"
* and even wider glob for -dbg:
  "*-dbg.rootfs.tar.bz2"

* replace them with the exact filename we expect for given image

* be aware that gzip won't accept the symlink in IMAGE_LINK_NAME causing:
  2023-03-13 08:58:23,845 - oe-selftest - INFO -  ... FAIL
  2023-03-13 08:58:23,845 - oe-selftest - INFO - Traceback (most recent call last):
    File "/OE/build/poky/meta/lib/oeqa/selftest/cases/imagefeatures.py", line 124, in test_bmap
      self.assertTrue(runCmd('gzip -t %s' % gzip_path))
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    File "/OE/build/poky/meta/lib/oeqa/utils/commands.py", line 214, in runCmd
      raise AssertionError("Command '%s' returned non-zero exit status %d:\n%s" % (command, result.status, exc_output))
  AssertionError: Command 'gzip -t /OE/build/poky/build/build-st-2023-03-12-todo-patch2/imagefeatures.ImageFeatures.test_bmap/build-st/tmp/deploy/images/qemux86-64/core-image-minimal-qemux86-64.rootfs.ext4.bmap.gz' returned non-zero exit status 1:
  gzip: skipping: /OE/build/poky/build/build-st-2023-03-12-todo-patch2/imagefeatures.ImageFeatures.test_bmap/build-st/tmp/deploy/images/qemux86-64/core-image-minimal-qemux86-64.rootfs.ext4.bmap.gz is a symbolic link

* and "*-dbg.rootfs.tar.bz2" doesn't work if IMAGE_NAME_SUFFIX
  is changed to anything else than ".rootfs" or moved into
  IMAGE_LINK_NAME (like I plan in future changes where this will
  need to be updated again).

  Luckily we can use the symlink which currently doesn't have
  IMAGE_NAME_SUFFIX nor the DATETIME, so we don't need to search
  for it with glob, e.g. currently:

  core-image-minimal-qemux86-64-dbg-20230313112546-dbg.rootfs.tar.bz2
  core-image-minimal-qemux86-64-dbg.tar.bz2 -> core-image-minimal-qemux86-64-dbg-20230313112546-dbg.rootfs.tar.bz2

[YOCTO #12937]

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
2 years agorunqemu: get_first_file() rename cmd* to glob*
Martin Jansa [Mon, 13 Mar 2023 12:15:40 +0000 (13:15 +0100)] 
runqemu: get_first_file() rename cmd* to glob*

* to better indicate how it's used in get_first_file

* cmd* is used in other places for actual shell commands
  to execute

* RunQemuError('KERNEL not found: %s, %s or %s' % cmds)
  also looked weird to me, but that works (to my python-noob surprise)

[YOCTO #12937]

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
2 years agoselftest: minidebuginfo.py respect IMAGE_LINK_NAME
Martin Jansa [Mon, 13 Mar 2023 12:15:38 +0000 (13:15 +0100)] 
selftest: minidebuginfo.py respect IMAGE_LINK_NAME

* use IMAGE_LINK_NAME instead of hardcoding
  core-image-minimal-${MACHINE} assumption

[YOCTO #12937]

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
2 years agoselftest: gdbserver.py: respect IMAGE_LINK_NAME
Martin Jansa [Mon, 13 Mar 2023 12:15:37 +0000 (13:15 +0100)] 
selftest: gdbserver.py: respect IMAGE_LINK_NAME

* use IMAGE_LINK_NAME instead of hardcoding
  core-image-minimal-${MACHINE} assumption

[YOCTO #12937]

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
2 years agoimage-artifact-names: add IMAGE_MACHINE_SUFFIX variable
Martin Jansa [Mon, 13 Mar 2023 12:15:36 +0000 (13:15 +0100)] 
image-artifact-names: add IMAGE_MACHINE_SUFFIX variable

* to make it easier for projects to avoid default -${MACHINE} suffix if
  the ${MACHINE} named DEPLOY_DIR_IMAGE works better for them

* also use IMAGE_LINK_NAME in IMAGE_NAME to make it more clear
  that IMAGE_NAME is the same as IMAGE_LINK_NAME but with version
  suffix

* adding it as separate variable helps us to catch the cases
  where we didn't respect ${IMAGE_LINK_NAME} variable and just used
  the common default ${IMAGE_BASENAME}-${MACHINE}.

[YOCTO #12937]

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
2 years agoselftest: fitimage.py: respect INITRAMFS_IMAGE_NAME and KERNEL_FIT_LINK_NAME
Martin Jansa [Mon, 13 Mar 2023 12:15:35 +0000 (13:15 +0100)] 
selftest: fitimage.py: respect INITRAMFS_IMAGE_NAME and KERNEL_FIT_LINK_NAME

* use these variables instead of assuming:
  "fitImage-its-%s-%s-%s" % (image_type, machine, machine)
  notice that these files have no extension and the -machine suffix
  is duplicated (once from core-image-minimal and once from
  kernel-fitimage.bbclass:kernel_do_deploy:append
  through KERNEL_FIT_LINK_NAME which is MACHINE by default:

  KERNEL_FIT_NAME ?= "${KERNEL_ARTIFACT_NAME}"
  KERNEL_FIT_LINK_NAME ?= "${KERNEL_ARTIFACT_LINK_NAME}"

  KERNEL_ARTIFACT_NAME ?= "${PKGE}-${PKGV}-${PKGR}-${MACHINE}${IMAGE_VERSION_SUFFIX}"
  KERNEL_ARTIFACT_LINK_NAME ?= "${MACHINE}"

[YOCTO #12937]

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
2 years agoimage-artifact-names.bbclass: add INITRAMFS_IMAGE_NAME from kernel.bbclass
Martin Jansa [Mon, 13 Mar 2023 12:15:34 +0000 (13:15 +0100)] 
image-artifact-names.bbclass: add INITRAMFS_IMAGE_NAME from kernel.bbclass

* move it from kernel.bbclass, because it needs to stay in sync with
  IMAGE_LINK_NAME structure

* image-artifact-names.bbclass is also inheritted from kernel-artifact-names.bbclass
  so every recipe which needs this variable probably already inherits one of these

* fixes kernel-fitimage.bbclass with modified IMAGE_LINK_NAME

[YOCTO #12937]

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
2 years agoselftest: runqemu: respect IMAGE_LINK_NAME
Martin Jansa [Mon, 13 Mar 2023 12:15:32 +0000 (13:15 +0100)] 
selftest: runqemu: respect IMAGE_LINK_NAME

* don't assume that every built image is named:

  <recipe>-<machine>.<suffix>

[YOCTO #12937]

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
2 years agoselftest: wic: respect IMAGE_LINK_NAME also in test_rawcopy_plugin_qemu
Martin Jansa [Mon, 13 Mar 2023 12:15:31 +0000 (13:15 +0100)] 
selftest: wic: respect IMAGE_LINK_NAME also in test_rawcopy_plugin_qemu

* this one is more tricky, because the test_rawcopy_plugin.wks.in file
  is used while building core-image-minimal-mtdutils, but the image filename
  inside wks.in is from core-image-minimal, so we cannot just let bitbake
  expand IMAGE_LINK_NAME, use separate variable set in the same config fragment
  IMAGE_LINK_NAME_CORE_IMAGE_MINIMAL

[YOCTO #12937]

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
2 years agoselftest: wic: respect IMAGE_LINK_NAME
Martin Jansa [Mon, 13 Mar 2023 12:15:30 +0000 (13:15 +0100)] 
selftest: wic: respect IMAGE_LINK_NAME

* use IMAGE_LINK_NAME instead of hardcoding
  core-image-minimal-${MACHINE} assumption

[YOCTO #12937]

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
2 years agoselftest: devtool: set BB_HASHSERVE_UPSTREAM when setting SSTATE_MIRROR
Martin Jansa [Mon, 13 Mar 2023 12:15:29 +0000 (13:15 +0100)] 
selftest: devtool: set BB_HASHSERVE_UPSTREAM when setting SSTATE_MIRROR

* with my build/conf/local.conf:
  SSTATE_DIR = "/OE/build/poky/build/sstate-cache"
  these devtool tests will first set own SSTATE_DIR and the original one set as SSTATE_MIRROR:

2023-03-11 11:51:46,837 - oe-selftest - INFO - test_devtool_update_recipe_append (devtool.DevtoolUpdateTests.test_devtool_update_recipe_append)
2023-03-11 11:51:46,846 - oe-selftest - DEBUG - Appending to: /OE/build/poky/build/build-st-2023-03-11-patch2/devtool.DevtoolUpdateTests.test_devtool_update_recipe_append/build-st/conf/selftest.inc
SSTATE_DIR = "/OE/build/poky/build/build-st-2023-03-11-patch2/devtool.DevtoolUpdateTests.test_devtool_update_recipe_append/build-st/sstate_devtool"
SSTATE_MIRRORS += "file://.* file:////OE/build/poky/build/sstate-cache/PATH"

* but that unfortunately leads to a warning from sanity.bbclass
  about SSTATE_MIRRORS without matching BB_HASHSERVE, because
  BB_HASHSERVE is set to "auto" by default

  these tests failing with:

2023-03-11 11:55:39,610 - oe-selftest - INFO - ======================================================================
2023-03-11 11:55:39,610 - oe-selftest - INFO - FAIL: test_devtool_update_recipe_append_git (devtool.DevtoolUpdateTests.test_devtool_update_recipe_append_git)
2023-03-11 11:55:39,610 - oe-selftest - INFO - ----------------------------------------------------------------------
2023-03-11 11:55:39,611 - oe-selftest - INFO - Traceback (most recent call last):
  File "/OE/build/poky/meta/lib/oeqa/selftest/cases/devtool.py", line 1118, in test_devtool_update_recipe_append_git
    self.assertNotIn('WARNING:', result.output)
AssertionError: 'WARNING:' unexpectedly found in 'NOTE: Starting bitbake server...\nWARNING: You are using a local hash equivalence server but have configured an sstate mirror. This will likely mean no sstate will match from the mirror. You may wish to disable the hash equivalence use (BB_HASHSERVE), or use a hash equivalence server alongside the sstate mirror.\nLoading cache...done.\nLoaded 0 entries from dependency cache.\nParsing recipes...done.\nParsing of 947 .bb files complete (0 cached, 947 parsed). 1764 targets, 52 skipped, 0 masked, 0 errors.\n\nSummary: There was 1 WARNING message.\nINFO: Updating SRCREV in recipe mtd-utils-selftest_git.bb\nNOTE: Writing append file /tmp/devtoolqa1m2lh02v/layer/recipes-devtools/mtd/mtd-utils-selftest_git.bbappend'
----------------------------------------------------------------------

* just setting BB_HASHSERVE to empty doesn't work, because then we
  would need to disable OEEquivHash as well as it fails with:

  ERROR: OEEquivHash requires BB_HASHSERVE to be set

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
2 years agotcf-agent: Update to current version
Mark Hatle [Mon, 20 Mar 2023 21:40:44 +0000 (16:40 -0500)] 
tcf-agent: Update to current version

While there has not been an official release in some time, the latest version
fixes a large number of bugs and adds support for Dwarf 5, among other
things.  (Dwarf 5 is default in gcc 12 and newer.)

Short change log since the last version:
    TCF Agent: Narrow down LoadLibrary search scope
    TCF Agent: faster breakpoint planting
    Bug 581214 - Expression evaluation fails for a data member that is defined in a structure, union or class that is of type DW_AT_data_bit_offset attribute (DWARF5)
    TCF Agent: reduced memory footprint
    TCF Agent: fixed build error on CygWin: CYGWIN_VERSION_CYGWIN_CONV undefined
    TCF Agent: workaround for GCC compilation error: function may return address of local variable
    Bug 581034 - "long long int" data type can't be recognized from dwarf5 debug info generated by llvm15
    TCF Agent: Removal of unused variable causing clang-15 build error
    TCF Agent: fixed: malformed HTTP request can crash the agent
    TCF Agent: fixed assertion failure in the Breakpoints service
    TCF Agent: GDB Remote Serial Protocol: fixed handling of process ID 0
    Bug 580644 - steps into the function of shared library does not work on PPC
    TCF Agent: add PowerPC 32-bit and big-endian builds in the Dockerfile
    TCF Agent: check glibc version when calling pthread_setname_np()
    Bug 580600 - -Wmisleading-indentation issue for gcc version less than 6.0
    Bug 580489 - failed to handle .debug_rnglists section sometimes
    Bug 580450 - failed to read the full compilation unit tag in dwarf 5 when customized sections
    Bug 580414 - failed to handle debug sections for DWARF 5
    Bug 580326 - Can not display the type of global variable defined in a sharedlib
    Fixed text formatting in Makefile.inc
    Bug 580279 - VERSION grep in agent/Makefile.inc need update for vxWorks
    Bug 580089 - pid2id() failed to get a right id
    TCF Agent: improved comments
    TCF Agent: a few more asserts
    TCF Agent: improved X86 disassembler
    Bug 580002 - [tcf-dev] Speedup compute_reverse_lookup_indices
    Bug 579989 - compile error by vs2008 on windows
    TCF Agent: improved X86 disassembler
    Bug 579947 - NULL pointer caused the tcf-server crashed when load ELF file that compiled by LLVM 14 with dwarf-5
    TCF Agent: fixed Coverity warning
    TCF Agent: fixed Coverity warning
    TCF Agent: new function in HTTP server API: closed() call-back
    Merge "TCF Agent: Fix potential memory/resource leaks"
    TCF Agent: Fix potential memory/resource leaks
    TCF Agent: Fix unsafe strcpy
    TCF Agent: Respect certain ENABLE_* macros
    Bug 579412 - incorrect or missing copyright information
    TCF Agent: Lazy PC initialization
    Bug 579362 - the process IDs should be pid_t instead of UINT32 in waitpid.c and waitpid.h files
    Bug 579378 - update copyright year to 2022
    Bug 579274 - unknown option -Wmisleading-indentation issue for gcc version less than 6.0
    TCF Agent: ARM v8 stack crawl: fixed handling of ERET instruction
    TCF Agent: fixed handling of line info file names when file contains a mix DWARF 5 and DWARF 3
    TCF Tests: fixed incorrect error reports in DWARF reader test
    TCF Agent: disabled bogus warnings misleading-indentation
    TCF Agent: fixed handling of line info file names in DWARF 5
    TCF Agent: Expressions: fixed handling of a static field of a struct stored in a register
    TCF Agent: fixed error when building with c++-11
    TCF Agent: DWARF 5: improved handling of location expressions
    TCF Agent: DWARF reader: fixed regression - "FORM_FLAG expected" error
    TCF Agent: DWARF reader: new object flag DOIF_inlined
    TCF Tests: improved tests of var declarations
    TCF Agent: fixed regression: memory corruption in the symbols proxy
    TCF Agent: fixed issues with handling GCC-11 debug info
    TCF Agent: fixed warning: ISO C90 forbids mixed declarations and code
    TCF Agent: Fix stack frame cleanup on error
    TCF Agent: Add missing parentheses
    Bug 578201 - Sometimes variables show as "N/A - Cannot read target memory. Input/output error"
    Bug 577936 - dprintf not work as expected
    TCF Agent: lazy initialization in rand32()
    TCF Agent: new API function: rand32()
    TCF Agent: fixed possible buffer overflow when calling fscanf()
    Bug 577174 - Sometimes variables show as "N/A - Value of register is unknown in the selected frame"
    Bug 577064 - Union type variables don’t show correctly if based on a register
    TCF Agent: fixed: tmp_vprintf and loc_vprintf can segfault on Linux
    Bug 577001 - DW_AT_high_pc with DW_FORM_udata form not handled
    TCF Agent: fixed pthread_cond_timedwait() on Windows
    TCF Agent: Expressions service: improved error message

Signed-off-by: Mark Hatle <mark.hatle@amd.com>
Signed-off-by: Mark Hatle <mark.hatle@kernel.crashing.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agomaintainers.inc: Modify email address
Zang Ruochen [Tue, 21 Mar 2023 11:06:09 +0000 (19:06 +0800)] 
maintainers.inc: Modify email address

zangrc.fnst@fujitsu.com -> zangruochen@loongson.cn

Signed-off-by: Zang Ruochen <zangruochen@loongson.cn>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopython3-jsonpointer: rewrite testing
Ross Burton [Tue, 21 Mar 2023 12:37:45 +0000 (12:37 +0000)] 
python3-jsonpointer: rewrite testing

The tests installed by jsonpointer use unittest not pytest, so update
the dependencies.

Use the unittest runner in python3-unittest-automake-output to get the
correct output.

Apply a patch so the test suite doesn't run twice.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopython3-webcolors: use python3-unittest-automake-output
Ross Burton [Tue, 21 Mar 2023 12:37:44 +0000 (12:37 +0000)] 
python3-webcolors: use python3-unittest-automake-output

Instead of using sed to reformat the output use this new module.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopython3-wcwidth: use python3-unittest-automake-output
Ross Burton [Tue, 21 Mar 2023 12:37:43 +0000 (12:37 +0000)] 
python3-wcwidth: use python3-unittest-automake-output

Instead of using sed to reformat the output use this new module.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopython3-pytz: use python3-unittest-automake-output
Ross Burton [Tue, 21 Mar 2023 12:37:42 +0000 (12:37 +0000)] 
python3-pytz: use python3-unittest-automake-output

Instead of using sed to reformat the output use this new module.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopython3-pyasn1: : use python3-unittest-automake-output
Ross Burton [Tue, 21 Mar 2023 12:37:41 +0000 (12:37 +0000)] 
python3-pyasn1: : use python3-unittest-automake-output

Instead of using sed to reformat the output use this new module.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopython3-pluggy: use python3-unittest-automake-output
Ross Burton [Tue, 21 Mar 2023 12:37:40 +0000 (12:37 +0000)] 
python3-pluggy: use python3-unittest-automake-output

Instead of using sed to reformat the output use this new module.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopython3-more-itertools: use python3-unittest-automake-output
Ross Burton [Tue, 21 Mar 2023 12:37:39 +0000 (12:37 +0000)] 
python3-more-itertools: use python3-unittest-automake-output

Instead of using sed to reformat the output use this new module. The
tests also need the statistics module so add this missing dependency.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopython3-markupsafe: use python3-unittest-automake-output
Ross Burton [Tue, 21 Mar 2023 12:37:38 +0000 (12:37 +0000)] 
python3-markupsafe: use python3-unittest-automake-output

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopython3-jinja2: use python3-unittest-automake-output
Ross Burton [Tue, 21 Mar 2023 12:37:37 +0000 (12:37 +0000)] 
python3-jinja2: use python3-unittest-automake-output

Instead of using sed to reformat the output use this new module.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopython3-hypothesis: use python3-unittest-automake-output
Ross Burton [Tue, 21 Mar 2023 12:37:36 +0000 (12:37 +0000)] 
python3-hypothesis: use python3-unittest-automake-output

Instead of using sed to reformat the output use this new module.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopython3-cryptography: use python3-unittest-automake-output
Ross Burton [Tue, 21 Mar 2023 12:37:35 +0000 (12:37 +0000)] 
python3-cryptography: use python3-unittest-automake-output

Instead of using sed to reformat the output use this new module.

This shows that the previous sed expression wasn't sufficient because it
didn't handle tests throwing an error, and now the bench/ tests fail
because we don't have pytest-benchmark.  These tests are more of a
benchmark than a unit test, so just skip them.

Also show a skip message instead of nothing if the tests were skipped
due to a lack of memory.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopython3-bcrypt: use python3-unittest-automake-output
Ross Burton [Tue, 21 Mar 2023 12:37:34 +0000 (12:37 +0000)] 
python3-bcrypt: use python3-unittest-automake-output

Instead of using sed to reformat the output use this new module.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopython3-atomicwrites: use python3-unittest-automake-output
Ross Burton [Tue, 21 Mar 2023 12:37:33 +0000 (12:37 +0000)] 
python3-atomicwrites: use python3-unittest-automake-output

Instead of using sed to reformat the output use this new module.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopython3-unittest-automake-output: add new recipe for ptest integration
Ross Burton [Tue, 21 Mar 2023 12:37:32 +0000 (12:37 +0000)] 
python3-unittest-automake-output: add new recipe for ptest integration

This package contains modules for both unittest and pytest that alter
the output to look like automake's 'make check' output, for better
integration with ptest.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopython3: add missing -modules dependencies
Ross Burton [Tue, 21 Mar 2023 12:37:31 +0000 (12:37 +0000)] 
python3: add missing -modules dependencies

The python3-modules package should depend on zoneinfo, and gdbm if it
was enabled.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopython3: missing ptest dependencies
Ross Burton [Tue, 21 Mar 2023 12:37:30 +0000 (12:37 +0000)] 
python3: missing ptest dependencies

The tests need a broad set of locales and timezones to be installed.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopython3-numpy: add missing dependency for the tests
Ross Burton [Tue, 21 Mar 2023 12:37:29 +0000 (12:37 +0000)] 
python3-numpy: add missing dependency for the tests

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agobinutils: Fix CVE-2023-25586
Yash Shinde [Thu, 16 Mar 2023 11:09:44 +0000 (16:39 +0530)] 
binutils: Fix CVE-2023-25586

Upstream-Status: Backport [https://sourceware.org/git/?p=binutils-gdb.git;h=5830876a0cca17bef3b2d54908928e72cca53502]

Signed-off-by: Yash Shinde <Yash.Shinde@windriver.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopython3-wheel: upgrade 0.38.4 -> 0.40.0
Tim Orling [Thu, 16 Mar 2023 01:03:01 +0000 (18:03 -0700)] 
python3-wheel: upgrade 0.38.4 -> 0.40.0

* Drop pyproject.toml patch, merged upstream

https://wheel.readthedocs.io/en/stable/news.html#release-notes

0.40.0 (2023-03-14)
* Added a wheel tags command to modify tags on an existing wheel (PR by
  Henry Schreiner)
* Updated vendored packaging to 23.0
* wheel unpack now preserves the executable attribute of extracted files
* Fixed spaces in platform names not being converted to underscores (PR
  by David Tucker)
* Fixed RECORD files in generated wheels missing the regular file attribute
* Fixed DeprecationWarning about the use of the deprecated pkg_resources
  API (PR by Thomas Grainger)
* Wheel now uses flit-core as a build backend (PR by Henry Schreiner)

License-Update: use LICENSE.txt for LIC_FILES_CHKSUM instead of PKG-INFO

Signed-off-by: Tim Orling <tim.orling@konsulko.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agosystemtap: Disable dangling-pointer warning
Khem Raj [Tue, 14 Mar 2023 18:00:04 +0000 (11:00 -0700)] 
systemtap: Disable dangling-pointer warning

This is to fix build in RISCV64

| In constructor 'symresolution_info::symresolution_info(systemtap_session&, bool)',
|     inlined from 'int semantic_pass_symbols(systemtap_session&)' at ../git/elaborate.cxx:1884:28:
| ../git/elaborate.cxx:2601:21: error: storing the address of local variable 'sym' in '*s.systemtap_session::symbol_resolver' [-Werror=dangling-pointer=]
|  2601 |   s.symbol_resolver = this; // save resolver for early PR25841 function resolution
|       |   ~~~~~~~~~~~~~~~~~~^~~~~~
| ../git/elaborate.cxx: In function 'int semantic_pass_symbols(systemtap_session&)':
| ../git/elaborate.cxx:1884:22: note: 'sym' declared here
|  1884 |   symresolution_info sym (s);
|       |                      ^~~
| ../git/elaborate.cxx:1884:22: note: 's' declared here

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agosystemd: fix wrong nobody-group assignment
Piotr Łobacz [Tue, 14 Mar 2023 14:17:50 +0000 (15:17 +0100)] 
systemd: fix wrong nobody-group assignment

The generated /etc/group file had a wrong group name for nobody-group
which was nobody with same id as nogroup groupd. This was leading to
duplcate groups, with same ids and different names.

More can be read on this link:
https://bugzilla.yoctoproject.org/show_bug.cgi?id=11766

Signed-off-by: Piotr Łobacz <p.lobacz@welotec.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agoresulttool: Improve overlapping ptest result reporting
Richard Purdie [Thu, 16 Mar 2023 22:51:10 +0000 (22:51 +0000)] 
resulttool: Improve overlapping ptest result reporting

With the move to a ptest per image, the multiple ptest log scenario
is much more likely. Tweak the handling to only warn if there are
overlapping files.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agoRevert "systemd-systemctl: Create machine-id with "uninitialized" text in it"
Chen Qi [Wed, 15 Mar 2023 10:11:04 +0000 (18:11 +0800)] 
Revert "systemd-systemctl: Create machine-id with "uninitialized" text in it"

This reverts commit 33d9b3a8ed186c2d212acb83f715367912010e7f.

The reverted patch was forcing systemd to do preset-all at boot time (first boot)
in a function that simuilates 'preset-all' at rootfs time. This is not right.

To force a first boot, just do not run this function, as said in the comment:
"""
    # If we populate the systemd links we also create /etc/machine-id, which
    # allows systemd to boot with the filesystem read-only before generating
    # a real value and then committing it back.
    #
    # For the stateless configuration, where /etc is generated at runtime
    # (for example on a tmpfs), this script shouldn't run at all and we
    # allow systemd to completely populate /etc.
"""

For example, removing 'systemd_preset_all;' from IMAGE_PREPROCESS_COMMAND in some
distro conf file will achieve the same effect of the original patch.

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agocargo.bbclass: use offline mode for building
Frederic Martinsons [Wed, 15 Mar 2023 07:48:51 +0000 (08:48 +0100)] 
cargo.bbclass: use offline mode for building

There should not be any network access during the build step so
specify this explicitely to cargo.
This will allow better error message, e.g:

| Caused by:
|  can't checkout from 'ssh://git@.../fmartinsons/zbus-git-dep-test.git': you are in  the offline mode (--offline)

Instead of

|   Updating git repository `ssh://git@.../fmartinsons/zbus-git-dep-test.git`
| warning: spurious network error (2 tries remaining): failed to resolve address for gitlab.com: Temporary failure in name resolution;class=Net (12)

Signed-off-by: Frederic Martinsons <frederic.martinsons@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agocracklib: update github branch to 'main'
Tim Orling [Thu, 16 Mar 2023 01:12:00 +0000 (18:12 -0700)] 
cracklib: update github branch to 'main'

Signed-off-by: Tim Orling <tim.orling@konsulko.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agomdadm: Fix raid0 tests
Mingli Yu [Thu, 16 Mar 2023 08:27:40 +0000 (16:27 +0800)] 
mdadm: Fix raid0 tests

Backport patches to fix raid0 tests:
  tests/00raid0
  tests/00readonly
  tests/03r0assem
  tests/04r0update
  tests/04update-metadata

Signed-off-by: Mingli Yu <mingli.yu@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agomdadm: fix tests/02lineargrow
Mingli Yu [Thu, 16 Mar 2023 08:27:39 +0000 (16:27 +0800)] 
mdadm: fix tests/02lineargrow

Backport patch [1] to fix tests/02lineargrow.

[1] https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/commit/?id=a2c832465fc75202e244327b2081231dfa974617

Signed-off-by: Mingli Yu <mingli.yu@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agomdadm: Fix testcase 06wrmostly
Mingli Yu [Thu, 16 Mar 2023 08:27:38 +0000 (16:27 +0800)] 
mdadm: Fix testcase 06wrmostly

Backport patch [1] to fix the failure of the 06wrmostly test.

[1] https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/commit/?id=41edf6f45895193f4a523cb0a08d639c9ff9ccc9

Signed-off-by: Mingli Yu <mingli.yu@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agoopenssl: Fix reproducibility issue
Richard Purdie [Thu, 16 Mar 2023 08:53:25 +0000 (08:53 +0000)] 
openssl: Fix reproducibility issue

Fix an issue introduced in the new openssl version where an assembler file
isn't generated in a reproducible way by seeding the perl random number
generator consistently. It has no crypto impact, it is just used to
avoid function name clashes.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agoopenssl: update from 3.0.8 to 3.1.0
Randy MacLeod [Wed, 15 Mar 2023 04:48:00 +0000 (00:48 -0400)] 
openssl: update from 3.0.8 to 3.1.0

>From the NEWS.md file:

  ### Major changes between OpenSSL 3.0 and OpenSSL 3.1.0 [14 Mar 2023]

  * SSL 3, TLS 1.0, TLS 1.1, and DTLS 1.0 only work at security level 0.
  * Performance enhancements and new platform support including new
    assembler code algorithm implementations.
  * Deprecated LHASH statistics functions.
  * FIPS 140-3 compliance changes.

Drop the upstreamed afalg.patch:
   c425e365f4 Configure: don't try to be clever when configuring afalgeng

Signed-off-by: Randy MacLeod <randy.macleod@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agomatchbox-terminal: Fix PV to match standard format
Richard Purdie [Thu, 16 Mar 2023 13:48:17 +0000 (13:48 +0000)] 
matchbox-terminal: Fix PV to match standard format

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopython3-urllib3: upgrade 1.26.14 -> 1.26.15
Wang Mingyu [Wed, 15 Mar 2023 11:13:34 +0000 (19:13 +0800)] 
python3-urllib3: upgrade 1.26.14 -> 1.26.15

Changelog:
==========
* Fix socket timeout value when "HTTPConnection" is reused ('#2645 <https://github.com/urllib3/urllib3/issues/2645>'__)
* Remove "!" character from the unreserved characters in IPv6 Zone ID parsing
  ('#2899 <https://github.com/urllib3/urllib3/issues/2899>'__)
* Fix IDNA handling of '\x80' byte ('#2901 <https://github.com/urllib3/urllib3/issues/2901>'__)

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopython3-testtools: upgrade 2.5.0 -> 2.6.0
Wang Mingyu [Wed, 15 Mar 2023 11:13:33 +0000 (19:13 +0800)] 
python3-testtools: upgrade 2.5.0 -> 2.6.0

Improvements
------------
* Add support for Python 3.10 and 3.11.
* Drop support for Python 3.5 (EOL).
* Distutils integration is deprecated and will be removed in the next major
  version.
* Use "CompoundFixture" from "fixtures>=2.0" rather than rolling our
  own.
* Provide a "testtools[twisted]" extra documenting dependencies needed for
  "testtools.twistedsupport".
* Make "TestCase" hashable.
* Prevent "AttributeError" in "TestCase.__eq__" for objects lacking a
  "__dict__" attribute.
* Replace deprecated "cgi" module usage with "email".

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopython3-scons: upgrade 4.4.0 -> 4.5.1
Wang Mingyu [Wed, 15 Mar 2023 11:11:23 +0000 (19:11 +0800)] 
python3-scons: upgrade 4.4.0 -> 4.5.1

Changelog:
==========
- Fix a problem in 4.5.0 where using something like the following code
  will cause a Clone()'d environment to share the CPPDEFINES with the
  original Environment() which was cloned. Causing leakage of changes
  to CPPDEFINES when they should be completely independent after the Clone.
      env=Environment(CPPDEFINES=['a'])
      env.Append(CPPDEFINES=['b']) (or AppendUnique,Prepend,PrependUnique)
      env1=env.Clone()
      env1.Append(CPPDEFINES=['c']) (or any other modification, but not overwriting CPPDEFINES
  Now env['CPPDEFINES'] will contain 'c' when it should not.
- Do not initialize DefaultEnvironment when calling EnsureSConsVersion(),
  EnsurePythonVersion(), Exit(), GetLaunchDir() and SConscriptChdir().
- Remove unused private method SConsEnvironment._exceeds_version().
- Added ValidateOptions() which will check that all command line options are in either
  those specified by SCons itself, or by AddOption() in SConstruct/SConscript.  It should
  not be called until all AddOption() calls are completed. Resolves Issue #4187
- Refactored SCons/Taskmaster into a package. Moved SCons/Jobs.py into that package.
  NOTE: If you hook into SCons.Jobs, you'll have to change that to use SCons.Taskmaster.Jobs
- Changed the Taskmaster trace logic to use python's logging module. The output formatting
  should remain (mostly) the same. Minor update to unittest for this to adjust for 1 less newline.
- Migrated logging logic for --taskmastertrace to use Python's logging module. Added logging
  to NewParallel Job class (Andrew Morrow's new parallel job implementation)
- Ninja: Fix execution environment sanitation for launching ninja. Previously if you set an
  execution environment variable set to a python list it would crash. Now it
  will create a string joining the list with os.pathsep
- Move execution environment sanitation from Action._subproc() to
  SCons.Util.sanitize_shell_env(ENV)
- Moved rpm and debian directories under packaging
- Added logic to help packagers enable reproducible builds into packaging/etc/. Please
  read packaging/etc/README.txt if you are interested.
- Added --experimental=tm_v2, which enables Andrew Morrow's new NewParallel Job implementation.
  This should scale much better for highly parallel builds. You can also enable this via SetOption().
- Fixed command line argument --diskcheck: previously a value of 'none' was ignored.
  SetOption('diskcheck','none') is unaffected, as it did not have the problem.
- Added overrides argument to SCons.Subst.scons_subst(), subst_list(), subst(), and Action's process(),
  strfunction(). This allows passing a dictionary of envvars to override when evaluating subst()'d strings/lists
- Fixed Issue #4275 - when outputting compilation db and TEMPFILE was in use, the compilation db would have
  command lines using the generated tempfile for long command lines, instead of the full command line for
  the compilation step for the source/target pair.
- Renamed the qt tools to qt3 since the logic in that tool is only for QT version 3.  Renamed all env vars
  which affect qt3 from QT_ to QT3_.  If you are still using SCons to build QT 3 code, you'll need to update
  your SConscripts.  Note that using 'qt' tool has been deprecated for some time.
- Added JAVAPROCESSORPATH construction variable which populates -processorpath.
- Updated JavaScanner to scan JAVAPROCESSORPATH.
- Added FILE_ENCODING, to allow explicitly setting the text encoding for files
  written by the Textfile() and Substfile() builders. If not specified, Textfile() and Substfile() builders
  will write files as UTF-8. Fixed Issue #4302.
- Add newline to end of compilation db (compile_commands.json).
- Added error message to handle the case when SCons attempts to retrieve all the targets
  for a specified builder from the CacheDir, fails to do so, and then runs into an error
  when deleting the files which were retrieved. Previously if this happened there was no
  errors or warnings.
- Fix issue #2757, where Configure checks that perform a check which reads a modified source
  (including program, source or header file(s)) would incorrectly mark that file "up to date" so the
  actual build would not see the file as modified. Leading to incorrect incremental builds.
  Now configure checks now clear node info for non conftest nodes, so they will be re-evaluated for
  the real taskmaster run when the build commences.
- Avoid returning UniqueList for `children` and other `Executor` APIs. This type
  iterates more slowly than the builtin types. Also simplify uniquer_hashables to
  use an faster implementation under the assumption of ordered dictionaries.
- Fixed runtest.py failure on Windows caused by excessive escaping of the path to python.exe.
- Run LaTeX after biber/bibtex only if necessary
- Added -fsanitize support to ParseFlags().  This will propagate to CCFLAGS and LINKFLAGS.
- A list argument as the source to the Copy() action function is now
  correctly handled by converting elements to string. Copy errors out
  if asked to copy a list to an existing non-directory destination.
  Both the implementation and the strfunction which prints the progress
  message were adjusted. Fixes #3009.
- doc: EnsureSConsVersion, EnsurePythonVersion, Exit, GetLaunchDir and
  SConscriptChdir are now listed as Global functions only; the
  Environment versions still work but are not documented.
- The Java scanner processing of JAVACLASSPATH for dependencies was
  changed to split on os.pathsep instead of space, to match usage of
  passing a path string like "xxx:yyy:zzz". This is not portable -
  passing a POSIX-style path string (with ':') won't work on Windows
  (';') - which is now documented with a hint to use a list instead
  to be portable. Splitting on space broke paths with embedded spaces.
  Fixes #4243.
- Cleanup: make sure BoolVariable usage in tests and examples uses Python
  boolean values instead of 0/1.
- Stop telling people to run "python setup.py install" in the User Guide.
  Adds new content on using virtualenvs to be able to have multiple
  different SCons versions available on one system.
- Added the "DefaultEnvironment(tools=[])" stanza to a number of tests
  that are known to be particularly slow.  It's still just a tiny
  speedup, but the Windows CI had been occasionally timing out,
  so maybe this helps a bit.
- Remove an extra existence check in one ninja test that caused it
  to be skipped on some otherwise-valid Windows installations.
- test framework tests now pass on Linux and Windows (the latter can
  still run into problems on some configurations), and automated
  tests are now run on changes in this area so future problems can
  be spotted.
- The single-file Util module was split into a package with a few
  functional areas getting their own files - Util.py had grown to
  over 2100 lines.
- Add a zipapp package of scons-local: can use SCons from a local
  file which does not need unpacking.
- Additional explanations for MSVSProject and MSVSSolution builders.
- Fix a problem (present in 4.4.0 only) where a Java inner class could
  not be cached because the emitted filename contained a '$' and when
  looked up through a node ended up generating a Python SyntaxError
  because it was passed through scons_subst().
- Have the gfortran tool do a better job of honoring user preferences
  for the dialect tools (F77, F90, F03 and F09, as well as the shared-library
  equivalents SHF77,  SHF90, SHF03, SHF09). Previously these were
  unconditionally overwritten to 'gfortran'; the change should be more
  in line with expectations of how these variables should work.
  Also cleaned a few Fortran tests - test behavior does not change.
- Updated MSVC documentation - adds "version added" annotations on recently
  added construction variables and provides a version-mapping table.
- Add Python 3.12 support, and indicate 3.11/3.12 support in package.
  3.12 is in alpha for this SCons release, the bytecode sequences
  embedded in SCons/ActionTests.py may need to change later, but
  based on what is known now, 3.12 itself should work with this release.
- Add "append" keyword argument to Configure context's CheckLib and
  CheckLibWithHeader to control whether to append or prepend (issue #2767)
  Also added "unique" keyword, to control whether a library is added
  or not if it is already in the $LIBS construction var in the
  configure context. (issue #2768).
- Completely refactored the CPPDEFINES logic in Append/AppendUnique/Prepend/PrependUnique
  This change fixes the following GH Issues:
  - GH Issue #3876 - Append() and AppendUnique() will handle CPPDEFINES the same
  - GH Issue #4254 - Make handling tuples in CPPDEFINES consistent.
  - We no longer sort the keys added to CPPDEFINES by their dictionary keys.
    We take advantage that their order is now stable based on insertion order
    in Python 3.5+
  - Added/modifed unit and system tests to verify these changes.

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopython3-pytest: upgrade 7.2.1 -> 7.2.2
Wang Mingyu [Wed, 15 Mar 2023 11:11:22 +0000 (19:11 +0800)] 
python3-pytest: upgrade 7.2.1 -> 7.2.2

Bug Fixes
=========
 #10533: Fixed pytest.approx() handling of dictionaries containing one or more values of 0.0.
 #10592: Fixed crash if --cache-show and --help are passed at the same time.
 #10597: Fixed bug where a fixture method named teardown would be called as part of nose teardown stage.
 #10626: Fixed crash if --fixtures and --help are passed at the same time.
 #10660: Fixed pytest.raises() to return a ‘ContextManager’ so that type-checkers could narrow pytest.raises(...) if ... else nullcontext() down to ‘ContextManager’ rather than ‘object’.

Improved Documentation
======================
 #10690: Added CI and BUILD_NUMBER environment variables to the documentation.
 #10721: Fixed entry-points declaration in the documentation example using Hatch.
 #10753: Changed wording of the module level skip to be very explicit about not collecting tests and not executing the rest of the module.

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopython3-poetry-core: upgrade 1.5.1 -> 1.5.2
Wang Mingyu [Wed, 15 Mar 2023 11:11:21 +0000 (19:11 +0800)] 
python3-poetry-core: upgrade 1.5.1 -> 1.5.2

Fixed
=======
 Fix an issue where wheels built on Windows could contain duplicate entries in the RECORD file (#555).

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopython3-setuptools: upgrade 67.4.0 -> 67.6.0
Wang Mingyu [Wed, 15 Mar 2023 11:11:20 +0000 (19:11 +0800)] 
python3-setuptools: upgrade 67.4.0 -> 67.6.0

Changelog:
=========
#3804: Added caching for supported wheel tags.
#3846: Added pruning heuristics to PackageFinder based on exclude.
#3843: Although pkg_resources has been discouraged for use, some projects still
consider pkg_resources viable for usage. This change makes it clear that
pkg_resources should not be used, emitting a DeprecationWarning when imported.

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agogrep: upgrade 3.8 -> 3.9
Wang Mingyu [Wed, 15 Mar 2023 11:11:19 +0000 (19:11 +0800)] 
grep: upgrade 3.8 -> 3.9

Bug fixes
==========
  With -P, some non-ASCII UTF8 characters were not recognized as
  word-constituent due to our omission of the PCRE2_UCP flag. E.g.,
  given f(){ echo Perú|LC_ALL=en_US.UTF-8 grep -Po "$1"; } and
  this command, echo $(f 'r\w'):$(f '.\b'), before it would print ":r".
  After the fix, it prints the correct results: "rú:ú".

  When given multiple patterns the last of which has a back-reference,
  grep no longer sometimes mistakenly matches lines in some cases.
  [Bug#36148#13 introduced in grep 3.4]

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agomatchbox-config-gtk: Update to latest SRCREV
Richard Purdie [Wed, 15 Mar 2023 17:58:40 +0000 (17:58 +0000)] 
matchbox-config-gtk: Update to latest SRCREV

License changes were the addition of SPDX license headers and license
header fixes.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agomatchbox-terminal: Update to latest SRCREV
Richard Purdie [Wed, 15 Mar 2023 17:57:49 +0000 (17:57 +0000)] 
matchbox-terminal: Update to latest SRCREV

License changes were the addition of SPDX license headers and license
header fixes.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agomatchbox-desktop-2: Update 2.2 -> 2.3
Richard Purdie [Wed, 15 Mar 2023 17:45:00 +0000 (17:45 +0000)] 
matchbox-desktop-2: Update 2.2 -> 2.3

License changes were the addition of SPDX license headers and license
header fixes.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agomatchbox-panel-2: Update 2.11 -> 2.12
Richard Purdie [Wed, 15 Mar 2023 17:40:39 +0000 (17:40 +0000)] 
matchbox-panel-2: Update 2.11 -> 2.12

License changes were the addition of SPDX license headers and license
header fixes.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agomatchbox-wm: Update 1.2.2 -> 1.2.3
Richard Purdie [Wed, 15 Mar 2023 17:24:44 +0000 (17:24 +0000)] 
matchbox-wm: Update 1.2.2 -> 1.2.3

Drop a patch merged in the upstream repo.
License changes are due to the addition of SPDX license headers.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopybootchart: Fix extents handling to account for cpu/io/mem pressure changes
Richard Purdie [Wed, 15 Mar 2023 14:42:38 +0000 (14:42 +0000)] 
pybootchart: Fix extents handling to account for cpu/io/mem pressure changes

The previous addition of pressure values to the chart didn't fix the extents
function which meant the bottom of the chart was cut off. Fix that.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopybootchartui: Fix python syntax issue
Richard Purdie [Wed, 15 Mar 2023 14:10:34 +0000 (14:10 +0000)] 
pybootchartui: Fix python syntax issue

Fix:

scripts/pybootchartgui/pybootchartgui/parsing.py:134: SyntaxWarning: "is" with a literal. Did you mean "=="?
  if pid is 0:

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agoscripts/combo-layer: Fix python deprecation warning
Richard Purdie [Tue, 14 Mar 2023 07:26:24 +0000 (07:26 +0000)] 
scripts/combo-layer: Fix python deprecation warning

Address:
DeprecationWarning: 'pipes' is deprecated and slated for removal in Python 3.13

pipes.quote is an alias for shlex.quote so switch to that.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agocpio: fix ptest failure
Changqing Li [Tue, 14 Mar 2023 02:13:01 +0000 (10:13 +0800)] 
cpio: fix ptest failure

Fix following ptest failure:
/usr/lib64/cpio/ptest/run-ptest: line 7: cd: /usr/lib/cpio/ptest/tests/: No such file or directory

Signed-off-by: Changqing Li <changqing.li@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agotzcode-native: fix build with gcc-13 on host
Martin Jansa [Mon, 13 Mar 2023 21:20:22 +0000 (21:20 +0000)] 
tzcode-native: fix build with gcc-13 on host

* passing -std=c2x to avoid build failure with gcc-13 on host
  works as well, but the resulting zic then segfaults when
  used in tzdata, use a fix from upstream instead

* reported upstream in https://mm.icann.org/pipermail/tz/2023-March/032690.html

* fixes:
  http://errors.yoctoproject.org/Errors/Details/697913/

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agogit-submodule-test: disable upstream version check
Martin Jansa [Mon, 13 Mar 2023 12:15:28 +0000 (13:15 +0100)] 
git-submodule-test: disable upstream version check

* this is one of the failures from distrodata.Distrodata.test_checkpkg:

  2023-03-11 14:26:21,482 - oe-selftest - INFO - ======================================================================
  2023-03-11 14:26:21,482 - oe-selftest - INFO - FAIL: test_checkpkg (distrodata.Distrodata.test_checkpkg)
  2023-03-11 14:26:21,482 - oe-selftest - INFO - ----------------------------------------------------------------------
  2023-03-11 14:26:21,482 - oe-selftest - INFO - Traceback (most recent call last):
    File "/OE/build/poky/meta/lib/oeqa/selftest/cases/distrodata.py", line 40, in test_checkpkg
      self.assertTrue(len(regressed_failures) == 0 and len(regressed_successes) == 0, msg)
  AssertionError: False is not true :
  The following packages failed upstream version checks. Please fix them using UPSTREAM_CHECK_URI/UPSTREAM_CHECK_REGEX
  (when using tarballs) or UPSTREAM_CHECK_GITTAGREGEX (when using git). If an upstream version check cannot be performed
  (for example, if upstream does not use git tags), you can set UPSTREAM_VERSION_UNKNOWN to '1' in the recipe to acknowledge
  that the check cannot be performed.
  git-submodule-test
  binutils

  Stdout:
  Loading cache...done.
  Loaded 0 entries from dependency cache.
  Parsing recipes...done.
  Parsing of 947 .bb files complete (0 cached, 947 parsed). 1764 targets, 46 skipped, 0 masked, 0 errors.
  ----------------------------------------------------------------------
  2023-03-11 14:26:21,482 - oe-selftest - INFO - Ran 1 test in 193.764s
  2023-03-11 14:26:21,482 - oe-selftest - INFO - FAILED
  2023-03-11 14:26:21,483 - oe-selftest - INFO -  (failures=1)
  2023-03-11 14:26:26,258 - oe-selftest - INFO - RESULTS:

  martin@jama /OE/build/poky/build $ devtool check-upgrade-status git-submodule-test
  NOTE: Could not list remote: Fetcher failure for URL: 'gitsm://git.yoctoproject.org/git-submodule-test;branch=master'.
    The command git -c gc.autoDetach=false -c core.pager=cat ls-remote git://git.yoctoproject.org/git-submodule-test refs/tags/* gave empty output unexpectedly
  INFO: git-submodule-test        1.0             UNKNOWN_BROKEN  None

  and there are no tags in this test repo:
  $ git ls-remote git://git.yoctoproject.org/git-submodule-test
  a2885dd7d25380d23627e7544b7bbb55014b16ee        HEAD
  d199bbf9ed2216bd1f38aec000d865ae08279119        refs/heads/changed-url
  a2885dd7d25380d23627e7544b7bbb55014b16ee        refs/heads/master
  049da4a6cb198d7c0302e9e8b243a1443cb809a7        refs/heads/ssh-gitsm-tests
  bbe99a1465c5cc52720936d075ddbc5ebe2ed731        refs/meta/cgit

* the binutils failure is strange, when I've added some debug output
  to see why it's listed the issue disappeared, maybe some temporary
  network glitch or something

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agotzdata: use separate B instead of WORKDIR for zic output
Martin Jansa [Mon, 13 Mar 2023 11:06:31 +0000 (12:06 +0100)] 
tzdata: use separate B instead of WORKDIR for zic output

* avoid copying whole exec_prefix over base_prefix as there
  were only zoneinfo files anyway

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agotimezone: use 'tz' subdir instead of ${WORKDIR} directly
Martin Jansa [Mon, 13 Mar 2023 11:06:30 +0000 (12:06 +0100)] 
timezone: use 'tz' subdir instead of ${WORKDIR} directly

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agoresulttool: add log --list-ptest
Ross Burton [Mon, 13 Mar 2023 12:59:08 +0000 (12:59 +0000)] 
resulttool: add log --list-ptest

Add a convenience argument to the log subcommand to list all of the
ptest logs in a testresults file.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agolib/resulttool: fix typo breaking resulttool log --ptest
Ross Burton [Mon, 13 Mar 2023 12:59:07 +0000 (12:59 +0000)] 
lib/resulttool: fix typo breaking resulttool log --ptest

ptestresult_get_log() looked for a key called 'ptestresuls.sections',
which should be 'ptestresult.sections'

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agovim: upgrade to 9.0.1403
Ross Burton [Mon, 13 Mar 2023 14:25:31 +0000 (14:25 +0000)] 
vim: upgrade to 9.0.1403

This incorporates fixes for CVE-2023-1127, CVE-2023-1170, CVE-2023-1175.

Also remove runtime/doc/uganda.txt from the license checksum: the Vim
license is also in the top-level LICENSE file so this is redundant.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agooeqa rtc.py: skip if read-only-rootfs
Mikko Rapeli [Mon, 13 Mar 2023 17:51:28 +0000 (19:51 +0200)] 
oeqa rtc.py: skip if read-only-rootfs

hwclock command fails on read-only-rootfs:

AssertionError: 1 != 0 : Failed to reset RTC time, output: hwclock: cannot open /etc/adjtime: Read-only file system

Signed-off-by: Mikko Rapeli <mikko.rapeli@linaro.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agovim: set modified-by to the recipe MAINTAINER
Ross Burton [Mon, 13 Mar 2023 16:48:09 +0000 (16:48 +0000)] 
vim: set modified-by to the recipe MAINTAINER

Clause II.3 of the Vim license states that any distribution of Vim that
has been extended or modified must _at least_ indicate in the :version
output that this is the case.

Handily, Vim has a --with-modified-by argument to add a line in that
text, so use MAINTAINER.  This is the distribution maintainer contact,
by default it is OE-Core Developers
<openembedded-core@lists.openembedded.org>.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agonpm.bbclass: avoid DeprecationWarning with new python
Martin Jansa [Mon, 13 Mar 2023 20:50:37 +0000 (21:50 +0100)] 
npm.bbclass: avoid DeprecationWarning with new python

meta/classes-recipe/npm.bbclass:85: DeprecationWarning: invalid escape sequence '\.'
  '--transform', 's,^\./,package/,',

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agofile: add few more PACKAGECONFIGs to avoid autodetected deps from host
Martin Jansa [Mon, 13 Mar 2023 19:37:59 +0000 (20:37 +0100)] 
file: add few more PACKAGECONFIGs to avoid autodetected deps from host

* there are few more autodetected compression libs
  --disable-zlib          disable zlib compression support [default=auto]
  --disable-bzlib         disable bz2lib compression support [default=auto]
  --disable-xzlib         disable liblzma/xz compression support
  --disable-zstdlib       disable zstdlib compression support [default=auto]
  --disable-lzlib         disable liblz (lzip) compression support

  and lzlib was detected on my gentoo (even when it doesn't say that default
  is auto detection)

  I've noticed this, because I've reused sstate-cache from my gentoo build
  in docker container with ubuntu-23.04 and esdk.oeSDKExtSelfTest.test_install_libraries_headers
  test was now failing with:

  Exception: subprocess.CalledProcessError: Command '/OE/build/poky/build/build-st-2023-03-12-esdk-patch1/esdk.oeSDKExtSelfTest.test_install_libraries_headers/build-st/tmp/work/x86_64-nativesdk-pokysdk-linux/buildtools-tarball/1.0-r0/recipe-sysroot-native/usr/bin/createrepo_c --update -q /OE/build/poky/build/build-st-2023-03-12-esdk-patch1/esdk.oeSDKExtSelfTest.test_install_libraries_headers/build-st/tmp/work/x86_64-nativesdk-pokysdk-linux/buildtools-tarball/1.0-r0/oe-sdk-repo' returned non-zero exit status 127.

  Subprocess output:
  /OE/build/poky/build/build-st-2023-03-12-esdk-patch1/esdk.oeSDKExtSelfTest.test_install_libraries_headers/build-st/tmp/work/x86_64-nativesdk-pokysdk-linux/buildtools-tarball/1.0-r0/recipe-sysroot-native/usr/bin/createrepo_c: error while loading shared libraries: liblz.so.1: cannot open shared object file: No such file or directory

  and liblz.so.1 ended in createrepo_c.real because of libmagic.so from file-native-replacement.

* use PACKAGECONFIG for seccomp as well and keep it disabled as it was
  explicitly disabled before (${@bb.utils.filter('DISTRO_FEATURES', 'seccomp', d)}
  can be added to default PACKAGECONFIG later if someone finds it useful)

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agosudo: update 1.9.12p2 -> 1.9.13p3
Xiangyu Chen [Mon, 13 Mar 2023 01:57:14 +0000 (09:57 +0800)] 
sudo: update 1.9.12p2 -> 1.9.13p3

Signed-off-by: Xiangyu Chen <xiangyu.chen@windriver.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agogo: use go as CVE product for all golang recipe veriants
Peter Marko [Sun, 12 Mar 2023 08:45:43 +0000 (09:45 +0100)] 
go: use go as CVE product for all golang recipe veriants

All golang vulnerabilities are reported under product 'go'.

By default there is no vulnerability reported for images with
golang components because none of used golang packages
have correct CVE product set:
* go-binary-native
* go-runtime
* go-cross-*

Signed-off-by: Peter Marko <peter.marko@siemens.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agorunqemu: add an option to enable guest-agent virtio device
Clément Péron [Fri, 10 Mar 2023 18:54:23 +0000 (19:54 +0100)] 
runqemu: add an option to enable guest-agent virtio device

Add support to the runqemu script for a new option, 'guestagent', that
enables the virtio serial port for host-to-guest communication.

Signed-off-by: Brenda Streiff <brenda.streiff@ni.com>
Signed-off-by: Clément Péron <peron.clem@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agoqemu: split out qemu-guest-agent, add startup scripts
Clément Péron [Fri, 10 Mar 2023 18:54:22 +0000 (19:54 +0100)] 
qemu: split out qemu-guest-agent, add startup scripts

Split out the QEMU guest agent into a separate package. The agent is
intended to be installed within a QEMU VM guest where a user is likely
to not want to have the rest of the QEMU installation within it.

Additionally, an initscript, udev rules file, and systemd unit file are
added to the package so that the guest agent can start automatically;
the former two come from Debian's packaging for qemu-guest-agent.

Signed-off-by: Brenda Streiff <brenda.streiff@ni.com>
Signed-off-by: Clément Péron <peron.clem@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agoimage_types: fix multiubi var init
Romuald Jeanne [Fri, 10 Mar 2023 10:36:21 +0000 (11:36 +0100)] 
image_types: fix multiubi var init

Make sure to expand all MKUBIFS_ARGS_<label> and UBINIZE_ARGS_<label> vars
in 'do_image_multiubi' task to use them to init the local 'mkubifs_args'
and 'ubinize_args' vars.

See [YOCTO #15065]

Signed-off-by: Romuald JEANNE <romuald.jeanne@st.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agosystemd.bbclass: Add /usr/lib/systemd to searchpaths as well
Khem Raj [Fri, 10 Mar 2023 10:07:03 +0000 (02:07 -0800)] 
systemd.bbclass: Add /usr/lib/systemd to searchpaths as well

Some packages like lirc places its unit files into $systemd_user_unitdir
and also uses them in SYSTEMD_SERVICE list in recipe. This fails in
do_package

ERROR: Didn't find service unit 'lircmd.service', specified in SYSTEMD_SERVICE:lirc.

here lircmd.service is installed in /usr/lib/systemd/system/lircmd.service

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agopypi.bbclass: Set SRC_URI downloadfilename with an optional prefix
Zoltan Boszormenyi [Fri, 10 Mar 2023 12:32:54 +0000 (13:32 +0100)] 
pypi.bbclass: Set SRC_URI downloadfilename with an optional prefix

This allows avoiding clashes between source archives of a main
project and a pypi project using the same name and version.

The new optional setting is PYPI_ARCHIVE_NAME_PREFIX which is empty
by default so previous downloads can be used. Example usage:

PYPI_ARCHIVE_NAME_PREFIX = "pypi-"

Signed-off-by: Zoltán Böszörményi <zboszor@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agodevtool/upgrade: do not delete the workspace/recipes directory
Alexander Kanavin [Thu, 9 Mar 2023 15:19:33 +0000 (16:19 +0100)] 
devtool/upgrade: do not delete the workspace/recipes directory

If it exists, there is no need to delete it, and if it does not,
devtool prints an ugly traceback.

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agoimage_types: fix vname var init in multiubi_mkfs() function
Romuald JEANNE [Thu, 9 Mar 2023 07:54:52 +0000 (08:54 +0100)] 
image_types: fix vname var init in multiubi_mkfs() function

As vname var is needed in multiubi_mkfs() function, we need to keep it
defined and use it as parameter to the new write_ubi_config() function.

See [YOCTO #15027]

Signed-off-by: Romuald JEANNE <romuald.jeanne@st.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agolibgit2: update license information
Sudip Mukherjee [Thu, 2 Mar 2023 21:47:44 +0000 (21:47 +0000)] 
libgit2: update license information

The LICENSE did not have complete information.
Some examples of missing license:

Zlib: deps/zlib/
ISC: tests/clar/clar.c
LGPL-2.1-or-later: src/libgit2/xdiff/xdiffi.c
CC0-1.0: src/util/rand.c

Signed-off-by: Sudip Mukherjee <sudipm.mukherjee@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agosystemd: Fix musl fix patch
Khem Raj [Sun, 12 Mar 2023 07:24:22 +0000 (23:24 -0800)] 
systemd: Fix musl fix patch

Current patch with fail when stdint.h is in include path before this
typedef which can also happen with re-includes. Therefore shunt that
by defining __DEFINED_wchar_t which helps musl builds of systemd-boot
recipe

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>
2 years agoptest-packagelists: Simplify ptest list/code
Richard Purdie [Sat, 11 Mar 2023 21:01:51 +0000 (21:01 +0000)] 
ptest-packagelists: Simplify ptest list/code

All the usage sites remove the -ptest suffix. Simply the original list
instead and clean up the code.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agocore-image-ptest: Switch to BBCLASSEXTEND parallel execution
Richard Purdie [Thu, 9 Mar 2023 22:00:08 +0000 (22:00 +0000)] 
core-image-ptest: Switch to BBCLASSEXTEND parallel execution

Replace the two images where tests are run serially with an image
per ptest in the form core-image-ptest-XXX using class extensions.

The original two images are then replaced with dependencies on these
images. This allows parallel execution of the tests.

The downside to this approach is the parsing time of an image generating
this number of class extensions is slow but making it easier to execute
and collect test data should outweigh that.

A useful advantage is that the dependencies of each ptest package are
checked/tested indvidually.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agobase-files: Drop localhost.localdomain from hosts file
Richard Purdie [Sat, 11 Mar 2023 07:50:45 +0000 (07:50 +0000)] 
base-files: Drop localhost.localdomain from hosts file

This was likely something we took inspiration from elsewhere with.
It was added in:

https://git.openembedded.org/openembedded/commit/packages/netbase/netbase/hosts?id=c8e5702127e507e82e6f68a4b8c546803accea9d

in 2005. Debian added this entry around 2004 and discussed and dropped
it in 2005:

https://lists.debian.org/debian-devel/2005/10/msg00559.html

resulting in:

https://salsa.debian.org/installer-team/netcfg/-/commit/3c15ee521b2b8f47b34ccc7f610523cd284f2221

We should drop this for some of the reasons in those threads,
it doesn't seem to be doing anything too helpful and isn't what most
applications expect.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agooeqa/selftest/sstate: Split classes to allow more parallelism
Richard Purdie [Sun, 12 Mar 2023 11:35:24 +0000 (11:35 +0000)] 
oeqa/selftest/sstate: Split classes to allow more parallelism

By splitting the single class into multiple classes, it allows more parallelism
in the execution. These tests don't have interdependencies that benefit
from sstate reuse or anything like that so this makes sense to improve overall
test execution time if/where possible.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agooeqa/selftest/sstate: Move common code to base class
Richard Purdie [Sun, 12 Mar 2023 11:31:38 +0000 (11:31 +0000)] 
oeqa/selftest/sstate: Move common code to base class

Move the other common shares test functions to the base class to
improve the code structure.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agooeqa/selftest/sstate: Merge sstate test class with tests themselves
Richard Purdie [Sun, 12 Mar 2023 11:29:27 +0000 (11:29 +0000)] 
oeqa/selftest/sstate: Merge sstate test class with tests themselves

Having this base class as a separate file is just confusing. Merge with
the rest of the test code.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 years agoselftest/recipetool: Stop test corrupting tinfoil class
Richard Purdie [Mon, 13 Mar 2023 11:44:09 +0000 (11:44 +0000)] 
selftest/recipetool: Stop test corrupting tinfoil class

Running recipetool.RecipetoolTests.test_recipetool_handle_license_vars
followed by wic.Wic2.test_biosplusefi_plugin_qemu would show a failure of:

  File "/media/build/poky/meta/lib/oeqa/utils/commands.py", line 351, in runqemu
    qemu = oeqa.targetcontrol.QemuTarget(recipedata, targetlogger, image_fstype)
  File "/media/build/poky/meta/lib/oeqa/targetcontrol.py", line 116, in __init__
    use_kvm = oe.types.qemu_use_kvm(d.getVar('QEMU_USE_KVM'), d.getVar('TARGET_ARCH'))
  File "/media/build/poky/meta/lib/oe/types.py", line 182, in qemu_use_kvm
    if kvm and boolean(kvm):
  File "/media/build/poky/meta/lib/oe/types.py", line 128, in boolean
    raise ValueError("Invalid boolean value '%s'" % value)
ValueError: Invalid boolean value '/media/build/poky/meta/files/common-licenses'

which made no sense until you realise the recipetool test is corrupting
the tinfoil class. Work on a copy instead to avoid this.

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