Changelog:
============
- New option '--bad-sector-data'.
- main_common.cc (format_num3): New function.
- mapbook.cc (input_pos_error): Print pos and size aligned.
- ddrescue.texi: Document use of -p and -x with --domain-mapfile.
- block.h: Rename to mapfile.h.
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
0001-Actually-use-the-hint-space-instruction-to-strip-PAC.patch
removed since it's inclued in 20250127.1
abseil-cpp/0004-abseil-ppc-fixes.patch
refreshed for 20250127.1
Changelog:
=============
- Added support for Bazel 8.0
- Added support for Bazel Platforms for better portability
- Added ABSL_ATTRIBUTE_VIEW and ABSL_ATTRIBUTE_OWNER for diagnosing certain lifetime issues
- Many performance improvements
- A security issue in hash container create/resize has been fixed. Note that the latest patch releases for previous LTS versions also address this issue.
- Bazel BUILD files now reference repositories by their canonical names from the Bazel Central Registry.
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Changelog
modernise CMakeLists.txt
Use cmake TIMESTAMP function to be more portable and to allow for reproducible builds
fixed: potential null deref. closes#556
python 3.13 compatibility
added: vs2022 support
fixed: creating (unsigned) windows installer
fixed: nsis installer not creating start menu icons and uninstaller
bump to .net framework 4.5 and .net 8.0
fixed: null terminate .net device language field. added CecLogicalAddresses::Unset()
fixed: don't call the config updated callback if nothing has changed. auto set wake devices or activate source so these values match what happens on the tv
fixed: tv would power on even if the options to power on were disabled in cec-tray
changed/fixed: windows installer, eventghost plugin, .net libraries, python imports, debug builds
fixed: use input() for python3 and raw_input() for python2
changed: debian python install to match arch specific too
fixed: libusb0.dll wasn't included in the installer
fixed: remove call to PyEcal_InitThreads(). swig already includes SWIG_PYTHON_INITIALIZE_THREADS in SWIG_init()
CheckPlatformSupport.cmake: various improvements
Ease copy-paste by invoking apt noninteractively
Rename COPYING to LICENSE.md so follow standard naming conversion
cecloader.h: fix null return
Mark all devices as inactive when new active source is unknown
Fix use after free in callbacks with results.
Add a commandHandler callback to allow external libs to handle CEC commands themselves.
Remove debug printfs.
Attempt to fix race condition when freeing the callback wrapper.
Allow command handler callback to intercept broadcast messages and fix inconsistent callback wrapper deletion again.
Remove useless mutex.
Remove useless 'keepResult' parameter and make sure the CCallbackWrap is only deleted once even in case of timeout.
Fix Python module import
fix assert issue with systemd udev
tegra-cec support. closes#636
Add support for Windows ARM64.
Signed-off-by: Markus Volk <f_l_k@t-online.de>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
0.30
----
Add support for configuring actions to be enabled or disabled over dbus:
- Actions can be listed via `powerprofilesctl list-actions`
- These can be controlled via `powerprofilesctl configure-action`
Add support for configuring whether PPD reacts to battery/AC charger events:
- It can be queried and controlled via:
`powerprofilesctl query-battery-aware` and
`powerprofilesctl configure-battery-aware`
Add better better compatibility with kernel 6.14.
0.23
----
Bugfix release for a performance regression with users who manually
adjusted AMD dGPU settings using other software.
0.22
----
Since this release power-profiles-daemon is also battery-level aware and
some drivers use this value to be smarter at tuning their optimizations.
In particular both the AMD panel power action now uses a progressive approach,
changing the the ABM based on the battery percentage.
AMD p-state received various features and improvements:
- it supports core performance boost when not in power-saver mode.
- uses minimum frequency to lowest non-linear frequency
- it is more impervious to faulty firmware and kernel bugs
Added support for changing DPM clocks on amdgpu, to explicitly set the DPM
clocks down to "low" when in power-saver mode.
The daemon command line interface now allows to disable logind and upower
integration.
Fixes handling of turbo_pct (now assumed as not present by default).
The systemd service lockdown settings have been restricted even more.
The systemd service starts later in the boot to avoid conflicts with
module loading.
0.21
----
Since this release power-profiles-daemon is battery-state aware and some drivers
use a more power efficient state when using the balanced profile on battery.
In particular both the AMD and Intel P-State drivers will use the
balance_power EPP profile, while for Intel one we also set the energy
performance bias to 8 (instead of 6).
This release also contains various fixes for the powerprofilesctl command line
tool when using the launch or version commands.
The tool is now better documented as we generate a manual page for it (if
python3-argparse is installed) and bash completions. We're even generating the
ZSH completions, but the install path must be provided.
The daemon command line interface has been improved for debugging, so use
--help-debug for further information.
The systemd service lockdown settings have been restricted even more.
Various code optimizations.
0.20
----
The project has moved under the freedesktop 'Upower' group. The service is
now advertised as 'org.freedesktop.UPower.PowerProfiles' in addition to the
previous 'net.hadess.PowerProfiles' for compatibility reasons.
This release adds support for:
* Multiple power-profiles-daemon drivers to load simultaneously. This notably
allows both CPU based control with amd-pstate or intel-pstate as well as
ACPI platform profile based control.
* amdgpu panel power savings which uses dedicated hardware in systems with
integrated Radeon graphics to decrease panel power consumption when the
system is on battery.
This release also enables the test suite by default, so distribution vendors
should update packaging accordingly.
0.13
----
This release adds support for the AMD P-State driver that's been added to the
6.3 Linux kernel. This release also fixes mismatched profiles on some HP
laptops and some miscellaneous bug fixes.
0.12
----
This release adds support for the Intel "Energy Performance Bias" feature, which
can be used on hardware that doesn't have a platform_profile or doesn't support
HWP. It will also be used to eke out a bit more performance, or power, on systems
which already supported HWP.
More information is available in the README.
0.11.1
------
This release stops power-profiles-daemon from modifying the cpufreq driver when
driver when the user/administrator has chosen to disable the Intel P-State scaling
governor (eg. forcing a passive operation mode).
More information is available in the README.
0.11
----
This release fixes problems on Intel machines when the CPUs didn't support turbo at
all, or the performance scaling governor was built as default in the kernel.
It also adds better end-user documentation, fixes in the command-line tool to not
cause bug report tools to popup on not-uncommon errors, and a bug fix for running
on some systems with controllable charge speeds.
0.10.1
------
This release fixes a bug in the authorization codepath added in 0.10.0, where holding
a profile could still succeed despite having sent a denial to the calling process.
This release also makes the pylint test optional. It should not be used unless the goal
is to contribute patches to assuage it.
0.10.0
------
This release adds authorization checks for the profile holds and profile switching
features of the backend daemon, through polkit. It is recommended that all
distributions upgrade to this version as soon as possible.
This release also adds support for the "quiet" kernel platform profile used
in some systems.
0.9.0
-----
This release adds support for "holding" a power profile while running a task
or application, making it possible to switch to a performance profile during
a compilation, or to a power-saver profile when low on battery, reverting to
the original profile when done.
This release also removes the "inhibited" property for the performance profile,
which made it impossible to switch to that profile, and replaces it with the
"degraded" property which lists why performance is degraded.
Finally, the last used profile is now remembered across reboots.
0.8.1
-----
This release works-around a cosmetic issue in gnome-shell animations when
on battery and using the intel-pstate driver.
0.8
Signed-off-by: Khem Raj <raj.khem@gmail.com>
This is a quick bugfix release that is API and ABI compatible with
previous 1.x releases.
- Handle SplitPCM wrong channels specifications. This fixes some
problems with disappearing devices.
- Add backwards compatibility support for when the kernel does not
support UMP. Also fix UMP output. This restores MIDI support for
older kernels/ALSA.
- Fix a crash in audioconvert because the resampler was not using the
right number of channels.
- Some compilation fixes and small improvements.
- Don't emit events when disconnecting a stream. (#3314)
- Fix some compilation problems. (#4603)
- Bump the ROC requirement to version 0.4.0
- Handle SplitPCM too few or too many channels. Add an error string
to the device names when the UCM config has an error.
- Add backwards compatibility support for when the kernel does not
support UMP.
- Configure the channels in the resampler correctly in all
cases. (#4595)
- Fix UMP output.
- Use the right samplerate of the filter-graph in audioconvert in
all cases.
- Fix a crash with an incomming call.
Signed-off-by: Markus Volk <f_l_k@t-online.de>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Upstream does not support our default parallel "make -j install"[0]. So,
disable it with PARALLEL_MAKEINST = "-j1" to avoid non-reproducibility
where some scripts can be installed or not.
Explanation for the non-reproducibility:
There is a race condition between 2 actions at install around the
installed script in $pkgdatadir:
* Removal of existing scripts /usr/share/poke/*.pk
* Installation of default scripts in the same directory
Sadly, those 2 actions are not ordered. Depending on the build system
load, removal can (rarely) happen *after* the installation. In this
case, no script in present in /usr/share/poke/ when the install process
end.
[0]: https://sourceware.org/bugzilla/show_bug.cgi?id=32815#c1
Signed-off-by: Yoann Congal <yoann.congal@smile.fr>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Adding to LDFLAGS works with LLD linker but not with BFD ld since it
gets added before the abseil-cpp libraries on linker cmdline which does
not link it and still finds the atomic function like `__atomic_store_8'
as missing
Use mipsarcho32 for override which covers both mips and mipsel
Thanks for suggestions - RAED [1]
[1] https://github.com/openembedded/meta-openembedded/pull/952
Signed-off-by: Khem Raj <raj.khem@gmail.com>
This fixes:
ERROR: pipewire-native-1.4.0-r0 do_populate_sysroot: sstate found an absolute path symlink /home/flk/poky/build/tmp/work/x86_64-linux/pipewire-native/1.4.0/sysroot-destdir/home/flk/poky/build/tmp/work/x86_64-linux/pipewire-native/1.4.0/recipe-sysroot-native/etc/alsa/conf.d/50-pipewire.conf pointing at /home/flk/poky/build/tmp/work/x86_64-linux/pipewire-native/1.4.0/recipe-sysroot-native/usr/share/alsa/alsa.conf.d/50-pipewire.conf. Please replace this with a relative link.
ERROR: pipewire-native-1.4.0-r0 do_populate_sysroot: sstate found an absolute path symlink /home/flk/poky/build/tmp/work/x86_64-linux/pipewire-native/1.4.0/sysroot-destdir/home/flk/poky/build/tmp/work/x86_64-linux/pipewire-native/1.4.0/recipe-sysroot-native/etc/alsa/conf.d/99-pipewire-default.conf pointing at /home/flk/poky/build/tmp/work/x86_64-linux/pipewire-native/1.4.0/recipe-sysroot-native/usr/share/alsa/alsa.conf.d/99-pipewire-default.conf. Please replace this with a relative link.
ERROR: pipewire-native-1.4.0-r0 do_populate_sysroot: Failing task due to absolute path symlinks
ERROR: Logfile of failure stored in: /home/flk/poky/build/tmp/work/x86_64-linux/pipewire-native/1.4.0/temp/log.do_populate_sysroot.913232
ERROR: Task (virtual:native:/home/flk/poky/meta-openembedded/meta-multimedia/recipes-multimedia/pipewire/pipewire_1.4.0.bb:do_populate_sysroot) failed with exit code '1'
Signed-off-by: Markus Volk <f_l_k@t-online.de>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Drop backported patches
Enable dependent options needed for thorvg packageconfig
Add packageconfig knob for lottie and its relationship with thorvg
Reviewed-by: Marek Vasut <marex@denx.de>
Cc: Marek Vasut <marex@denx.de>
Cc: Jaeyoon Jung <jaeyoon.jung@lge.com>
Cc: Martin Jansa <martin.jansa@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Upgrade to release 2.13.6:
- A new self._pybind11_conduit_v1_() method is automatically added
to all py::class_-wrapped types, to enable type-safe
interoperability between different independent Python/C++
bindings systems, including pybind11 versions with different
PYBIND11_INTERNALS_VERSION's. Supported on pybind11 2.11.2,
2.12.1, and 2.13.6+.
- Using __cpp_nontype_template_args instead of
__cpp_nontype_template_parameter_class.
- Properly translate C++ exception to Python exception when
creating Python buffer from wrapped object.
- Adds an answer (FAQ) for "What is a highly conclusive and simple
way to find memory leaks?".
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
rv32 does not have compiler builtins for 64bit atomics
Fixes
| riscv32-yoe-linux-ld.lld: error: undefined reference: __atomic_load_8
| >>> referenced by /mnt/b/yoe/master/build/tmp/work/riscv32-yoe-linux/protobuf/5.29.4/recipe-sysroot/usr/lib/libabsl_cordz_info.so.2501.0.0 (disallowed by --no-allow-shlib-undefined)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
This allows the default do_configure to run, which will run "make clean"
if needed.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
* some recipes which use googletest fail with:
gtest/src/gtest_main.cc:32: recipe-sysroot/usr/include/c++/15.0.1/ciso646:46:4: error: #warning "<ciso646> is deprecated in C++17, use <version> to detect implementation-specific macros" [-Werror=cpp]
46 | # warning "<ciso646> is deprecated in C++17, use <version> to detect implementation-specific macros"
| ^~~~~~~
backport a fix for this.
* see https://github.com/google/googletest/issues/4701
Signed-off-by: Martin Jansa <martin.jansa@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
* fixes:
http://errors.yoctoproject.org/Errors/Details/848488/
which is triggered only when thorvg PACKAGECONFIG is enabled
Signed-off-by: Martin Jansa <martin.jansa@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
* backport fix from:
690342b9cb
to fix:
http://errors.yoctoproject.org/Errors/Details/848457/
grammar.y:58:33: error: too many arguments to function 'libconfig_yyget_lineno'; expected 0, have 1
* need to add bison-native dependency while grammar.y is modified
with the above patch to fix:
http://errors.yoctoproject.org/Errors/Details/848472/
../aux-build/ylwrap: line 174: yacc: command not found
Signed-off-by: Martin Jansa <martin.jansa@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Catch more CVEs by adding additional product per
sqlite> select vendor, product, count(*) from products where product like '%libmad%' group by vendor, product;
underbit|libmad|1
underbit|mad_libmad|4
Signed-off-by: Peter Marko <peter.marko@siemens.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
CPE does not contain mnajor version number, so set VE product to just
freerdp.
Without this there are no (fixed) CVEs in reports.
Signed-off-by: Peter Marko <peter.marko@siemens.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Upgrade to release 3.3.6:
- Fix a false positive for used-before-assignment when an inner
function's return type annotation is a class defined at module
scope.
- Fix false positives for use-implicit-booleaness-not-comparison,
use-implicit-booleaness-not-comparison-to-string and
use-implicit-booleaness-not-comparison-to-zero when chained
comparisons are checked.
- Fix a false positive for invalid-getnewargs-ex-returned when the
tuple or dict has been assigned to a name.
- Remove getopt and optparse from the list of deprecated modules.
- Fixed conditional import x.y causing false positive
possibly-used-before-assignment.
- Fix a crash when something besides a class is found in an except
handler.
- Fixed raising invalid-name when using camelCase for private
methods with two leading underscores.
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
