Set UPSTREAM_CHECK_REGEX to check the correct latest stable verison.
Before the fix:
$ devtool latest-version nmap
INFO: Current version: 7.95
INFO: Latest version: 7.95-1
After the fix:
$ devtool latest-version nmap
INFO: Current version: 7.95
INFO: Latest version: 7.95
Signed-off-by: Yi Zhao <yi.zhao@windriver.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Backport new API signature, since the vendored version of openssl in
libspdm is 3.0.14, asn1.h has to sync for providing it.
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Cc: Alistair Francis <alistair.francis@wdc.com>
Cc: Martin Jansa <martin.jansa@gmail.com>
Avoid following warning by using a dedicated unpack directory for sources.
WARNING: kernel-selftest-1.0-r0 do_unpack: kernel-selftest: the directory ${WORKDIR}/${BP} (/home/kraj200/yoe/build/tmp/work/qemux86_64-yoe-linux/kernel-self test/1.0/kernel-selftest-1.0) pointed to by the S variable doesn't exist - please set S within the recipe to point to where the source has been unpacked to
Signed-off-by: Khem Raj <raj.khem@gmail.com>
The upstream code basically hard-code libdir to /usr/lib. This was
hidden by the upstream non-reproducible path existence check.
Improve the reproducibility patch to use the cmake variables derived
from $libdir and $datadir :
CMAKE_INSTALL_LIBDIR and CMAKE_INSTALL_DATAROOTDIR.
Signed-off-by: Yoann Congal <yoann.congal@smile.fr>
Reported-by: Martin Jansa <martin.jansa@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
This recipe is needed to build latest version of matplotlib python module
when internal vendored modules are not used which is preferred here, avoids
complicating recipe to counter meson's wrap logic
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Cc: Rob Woolley <rob.woolley@windriver.com>
Latest OpenCL CTS is released after the latest OpenCL Headers and uses a
symbol which is not defined in the release. Backport a patch to make
OpenCL CTS compile.
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Upgrade OpenCL-Headers to the latest release. Upgrade opencl-clhpp with
the same commit so that they stay in sync, otherwise projects using
clhpp might fail to build.
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
After upgrading GCC—for example, from 14.1.0 to 14.2.0—building lmsensors that
was previously compiled with GCC 14.1.0 may fail with an error like:
lmsensors/3.6.0/recipe-sysroot-native/usr/lib/x86_64-wrs-linux/gcc/x86_64-wrs-linux/
14.1.0/include/stddef.h can't find, which is needed by 'prog/sensord/args.rd'.
This occurs because prog/sensord/args.rd still references stale headers from the
older GCC version.
The root cause is that stale *.rd and *.ro files under prog/sensord are not
properly cleaned during do_configure. This patch ensures those files are removed
to prevent broken dependencies when GCC is upgraded.
Also remove the same statement in do_compile.
Signed-off-by: Haixiao Yan <haixiao.yan.cn@windriver.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Remove this not maintained recipe. The latest commit of pmtools is 12
years ago. And it generates 3 binaries: acpidump/acpixtract/madt, but
there is a recipe acpica in oe-core provides acpidump/acpixtract, and
according to README.madt, it is hacked out from Linux kernel, if someone
requests it, mayb user can add one recipe similar like turbostat.
[1] https://github.com/anyc/pmtools/tree/master
Signed-off-by: Changqing Li <changqing.li@windriver.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
* fix too many arguments error
dhry_1.c: In function 'main':
dhry_1.c:176:23: error: too many arguments to function 'Func_1'; expected 0, have 2
176 | if (Enum_Loc == Func_1 (Ch_Index, 'C'))
| ^~~~~~ ~~~~~~~~
dhry_1.c:33:17: note: declared here
33 | Enumeration Func_1 ();
| ^~~~~~
Signed-off-by: mark.yang <mark.yang@lge.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
* see more details
http://errors.yoctoproject.org/Errors/Details/851798/
lat_rpc.c:172:1: error: conflicting types for 'client_rpc_xact_1'; have 'char *(char *, CLIENT *)' {aka 'char *(char *, struct __rpc_client *)'}
172 | client_rpc_xact_1(char *argp, CLIENT *clnt)
| ^~~~~~~~~~~~~~~~~
bench.h:349:14: note: previous declaration of 'client_rpc_xact_1' with type 'char *(void)'
349 | extern char *client_rpc_xact_1();
| ^~~~~~~~~~~~~~~~~
lat_rpc.c: In function 'rpc_xact_1':
lat_rpc.c:189:1: warning: old-style function definition [-Wold-style-definition]
189 | rpc_xact_1(msg, transp)
| ^~~~~~~~~~
lat_rpc.c:192:1: error: number of arguments doesn't match prototype
192 | {
| ^
bench.h:348:14: error: prototype declaration
348 | extern char *rpc_xact_1();
| ^~~~~~~~~~
Fix errors due to old-style function declarations
The code was using old-style function declarations without proper prototypes,
which causes compilation errors with newer GCC versions. This patch updates
the function declarations to use modern C syntax with proper parameter types.
Signed-off-by: mark.yang <mark.yang@lge.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
A valid ELF file may contain a SHT_NOBITS .BTF section. This case is not
handled correctly in btf_parse_elf, which leads to a segfault.
Signed-off-by: Changqing Li <changqing.li@windriver.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
* Fix the following build error with gcc-15:
nbench1.c:3520:13: error: conflicting types for 'adjust_mid_wts'; have 'void(int)'
3520 | static void adjust_mid_wts(int patt)
| ^~~~~~~~~~~~~~
In file included from nbench1.c:64:
nbench1.h:373:13: note: previous declaration of 'adjust_mid_wts' with type 'void(void)'
373 | static void adjust_mid_wts();
| ^~~~~~~~~~~~~~
Signed-off-by: mark.yang <mark.yang@lge.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Use PACKAGECONFIG to support keyutils and openssl by default. This
allows NVMe-OF connections with TLS.
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Add a recipe for b4 - a utility for working with mailing-list based
development workflows.
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
valgrind does not build for RISCV platforms yet and so are some other
platforms ( taken from rsyslog recipe )
Signed-off-by: Khem Raj <raj.khem@gmail.com>
* gcc-15 uses gnu23 for c.
'bool' is a keyword.
TOPDIR/tmp/work/core2-64-oe-linux/paho-mqtt-c/1.3.14/git/src/MQTTPacket.h:31:22: error: 'bool' cannot be defined via 'typedef'
31 | typedef unsigned int bool;
| ^~~~
TOPDIR/tmp/work/core2-64-oe-linux/paho-mqtt-c/1.3.14/git/src/MQTTPacket.h:31:22: note: 'bool' is a keyword with '-std=c23' onwards
TOPDIR/tmp/work/core2-64-oe-linux/paho-mqtt-c/1.3.14/git/src/MQTTPacket.h:31:1: warning: useless type name in empty declaration
31 | typedef unsigned int bool;
| ^~~~~~~
Signed-off-by: mark.yang <mark.yang@lge.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
* Backport fix from:
https://github.com/google/perfetto/commit/3953f56f98420e2ecb0e1c4542e20bfbb81da965
* To fix the error with gcc-15.
http://errors.yoctoproject.org/Errors/Details/851189/
../git/include/perfetto/ext/tracing/core/slice.h:47:46: error: 'uint8_t' was not declared in this scope
47 | static Slice TakeOwnership(std::unique_ptr<uint8_t[]> buf, size_t size) {
| ^~~~~~~
../git/include/perfetto/ext/tracing/core/slice.h:25:1: note: 'uint8_t' is defined in header '<cstdint>'; this is probably fixable by adding '#include <cstdint>'
24 | #include <string>
+++ |+#include <cstdint>
Signed-off-by: mark.yang <mark.yang@lge.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
ver 3.6:
Fix issue with handling blacklisting and roaming requests.
Fix issue with handling CQM thresholds for FullMAC devices.
Add support for PMKSA when using FullMAC devices.
ver 3.5:
Add support for option to disable blacklist handling.
Add support for option to disable SAE for broken drivers.
ver 3.4:
Add support for the Test Anything Protocol.
Signed-off-by: Markus Volk <f_l_k@t-online.de>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Package ktls-utils which includes tlsd.
This is used when in-kernel TLS consumers need a mechanism to perform TLS
handshakes on a connected socket to negotiate TLS session parameters that
can then be programmed into the kernel's TLS record protocol engine.
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
* A strict error handling for function parameters in gcc-15 causes errors.
output.c:9:6: error: conflicting types for 'initcolors'; have 'void(tOutput *)' {aka 'void(struct _tOutput *)'}
9 | void initcolors(tOutput* output)
| ^~~~~~~~~~
In file included from output.c:7:
output.h:10:6: note: previous declaration of 'initcolors' with type 'void(void)'
10 | void initcolors();
| ^~~~~~~~~~
Signed-off-by: mark.yang <mark.yang@lge.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
* Imported the submitted patch from:
https://savannah.gnu.org/bugs/?66845
to fix:
http://errors.yoctoproject.org/Errors/Details/851187/
../../enscript-1.6.6/compat/regex.c:3565:13: error: too many arguments to function 're_match_2_internal'; expected 0, have 8
3565 | val = re_match_2_internal (bufp, string1, size1, string2, size2,
| ^~~~~~~~~~~~~~~~~~~ ~~~~
* Modified the Upstream-Status since the same content as
0001-getopt-Include-string.h-for-strcmp-stcncmp-functions.patch is already in the backport.
Signed-off-by: mark.yang <mark.yang@lge.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
* Fix prototype declaration and too many arguments function errors due to outdated style
byte_chr.c:9:1: error: number of arguments doesn't match prototype
9 | {
| ^
In file included from byte_chr.c:3:
byte.h:6:21: error: prototype declaration
6 | extern unsigned int byte_chr();
| ^~~~~~~~
...
./compile buffer_get.c
buffer_get.c: In function 'oneread':
buffer_get.c:12:9: error: too many arguments to function 'op'; expected 0, have 3
12 | r = op(fd,buf,len);
| ^~ ~~
...
supervise.c: In function 'doit':
supervise.c:144:11: error: too many arguments to function 'wait_nohang'; expected 0, have 1
144 | r = wait_nohang(&wstat);
| ^~~~~~~~~~~ ~~~~~~
Signed-off-by: mark.yang <mark.yang@lge.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
OpenOCD non reproducible build has been fixed.
See commit : 60d88818646dc418c3b505ea8b2e24c9a7e48ef0
Remove OpenOCD from the known non reproducibles.
Signed-off-by: Sofiane HAMAM <sofiane.hamam@smile.fr>
Signed-off-by: Yoann Congal <yoann.congal@smile.fr>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Main problem with ptest for this component is that it uses cmake's
FetchContent and CPM to get dependencies.
This adds lot of ugly code to the recipe including conditional patch.
Second big problem is that tests need BIT7Z_DISABLE_USE_STD_FILESYSTEM
which uses test library. This means that when building with ptests, the
code is significantly different than when building without it.
But in production case we don't want to use testing library...
This is known at upstream and will be fixed eventually as github CI is
failing on this too when submitting unrelated patches upstream.
Other considerations:
* created patch for new cmake option to pass path to test data on target
* created patch for new cmake option to pass path to lib7zip on target
* skipped test which consumes too much RAM (it passes if machine has
plenty of RAM)
* testdata contains files for other architectures, so INSANE_SKIP is
needed for ptest package
* created patch for tests failing with musl
Tests usually take 9s on my build machine so added them to fast ptests.
However since the dependency 7zip recipe does not build on 64-bit
architectures, I could not add it to PTESTS_FAST_META_OE.
Signed-off-by: Khem Raj <raj.khem@gmail.com>
bit7z is a cross-platform C++ static library that allows the
compression/extraction of archive files through a clean and simple
wrapper interface to the dynamic libraries from the 7-Zip project.
It supports compression and extraction to and from the filesystem or
the memory, reading archives metadata, updating existing ones,
creating multi-volume archives, operation progress callbacks, and many
other functionalities.
Recipe comments:
* 2 patches needed for successful build+ptest were submitted upstream
* to upstream dependency inclusion patch we'd have to completely rework
dependency handling and would be probably against their concepts
Signed-off-by: Peter Marko <peter.marko@siemens.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Jeroen Hofstee
Khem Raj
Yi Zhao
Jason Schonberg
Yoann Congal
Etienne Cordonnier
Dmitry Baryshkov
Haixiao Yan
Changqing Li
mark.yang
Alistair Francis
Bartosz Golaszewski
Markus Volk
Wang Mingyu
Martin Jansa
Sofiane HAMAM
Peter Marko