Maximum number of tries, in rare cases, is insufficient for
elf parse. Backport patch that fixes the issue.
Signed-off-by: Sakib Sajal <sakib.sajal@windriver.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
The patch to remove these options is removed therefore instrument them
via sed operations
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Cc: Mingli Yu <mingli.yu@windriver.com>
Newer versions of compilers get stricter and add new diagnostics which
cause build failures that upstream should address not distros
Fixes
| In function 'perf_table_alloc_umask',
| inlined from 'gen_tracepoint_table' at pfmlib_perf_event_pmu.c:434:10:
| pfmlib_perf_event_pmu.c:304:47: error: pointer may be used after 'realloc' [-Werror=use-after-free]
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Userspace tool from the kernel. Used for control the client/server of
USB/IP. The server expect USB devices and the client import them on the
client machine. Making it look like a device is physical connected to
the client, even though it is physical connected to the server.
Signed-off-by: Claus Stovgaard <clst@ambu.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
We generally want to use ipmitool without the event daemon. Generally
this utility is less commonly used and takes up space in our images.
Signed-off-by: William A. Kennington III <wak@google.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
* Drop 0002-mem_section-Support-only-46-bit-for-MAX_PHYSMEM_BITS.patch
This patch has been introduced in January 2017 for makedumpfile 1.6.1.
It has never reached upstream, hence the pending status is
inappropriate. Since that time, the upstream code to retrieve
MAX_PHYSMEM_BITS has been updated to use vmcoreinfo, with a backward
compatibility for older kernels.
* Remove the deprecated comment about arm/aarch64 would compile but has never
been tested upstream, introduced in July 2016. Nowadays, it works just fine
on those architectures.
* Add PACKAGECONFIG option for zstd compression
Since 1.7.0 release, Zstandard (zstd) compression support has been added.
With zstd, the vmcore size and time consumption can have a better balance
than zlib/lzo/snappy.
http://lists.infradead.org/pipermail/kexec/2021-September/023011.html
Signed-off-by: Fathi Boudra <fathi.boudra@linaro.org>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
This patch updates SRC_URIs using git to include branch=master if no branch is set
and also to use protocol=https for github urls as generated by the conversion script
in OE-Core.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Refresh the following patch:
donnot-extract-gdb-during-do-compile.patch
remove-unrecognized-gcc-option-m32-for-mips.patch
0001-printk-add-support-for-lockless-ringbuffer.patch
0002-printk-use-committed-finalized-state-values.patch
Removed since these are included in 7.3.0.
Signed-off-by: Zang Ruochen <zangrc.fnst@fujitsu.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
This dependency introduces a cycle with pahole:
```
kernel --> pahole --> libbpf
^ /
\ /
`------------------<
```
pahole is a build time dependency of the kernel which is required to
generate BTF symbols. The dependency on the kernel configuration is nice
to have but makes using pahole during kernel compile time impossible.
Furthermore this dependency doesn't make sense when building libbpf
purely to ship it via a SDK.
Signed-off-by: Peter Morrow <pemorrow@linux.microsoft.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
This is the result of automated script (0.9.1) conversion:
oe-core/scripts/contrib/convert-overrides.py .
converting the metadata to use ":" as the override character instead of "_".
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
libpfm4 is only enabled for powerpc arch as of now.
This enables the lib on Arm 64bit platform as well.
Signed-off-by: Olivier Georget <olivier.georget@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
The license files were renamed in oe-core to match the SPDX names.
Most recipes here were already updated in commit ed54f12e19
("recipes: Update common-licenses references to match new names"),
but spidev-test was still missing.
Signed-off-by: Daniel Klauer <daniel.klauer@gin.de>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
if cpu exec line was split into different line, there will be an error
when parse it:
File "/usr/bin/sleepgraph", line 3165, in parseTraceLog
proclist[name] = int(val[1])
IndexError: list index out of range
check this case and append to parse the unfinished line.
Signed-off-by: Liwei Song <liwei.song@windriver.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
perf(oe-core) also uses the doc included in plugins/, so package it in own subdirs of trace-cmd.
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Pass -f(debug|macro)-prefix-map through CC because libbpf
build system overrides CFLAGS. Also remap STAGING_KERNEL_DIR
because it is also stored in resulting executable
Signed-off-by: Denys Zagorui <dzagorui@cisco.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Have not seen build issues for 5.4 and 5.10
x86-64 kernels so allow to build it
Signed-off-by: Denys Zagorui <dzagorui@cisco.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
its checking for .config values which means we need to stage kernel
sources before executing do_compile
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Cc: Matteo Croce <mcroce@microsoft.com>
The libbpf makefile uses pkg-config to get the libelf build flags and
file paths.
Inherit pkgconfig so the install target can copy the binaries in the
sysroot, which are needed by other packages, like bcc.
Signed-off-by: Matteo Croce <mcroce@microsoft.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
with 5.10 it builds bpf-bootstrap tool and then tries to run it as well
which wont work in cross builds. Disable it until it can be make cross
compile friendly.
Signed-off-by: Khem Raj <raj.khem@gmail.com>
crash/0001-Fix-for-an-ARM64-gcc-10-compilation-error.-Without-t.patch
Removed since this is included in 7.2.9
Signed-off-by: Zheng Ruoqin <zhengrq.fnst@cn.fujitsu.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
* Starting with 1.6.8 makedumpfile project migrated from sourceforge to github;
properly updated SRC_URI, HOMEPAGE, UPSTREAM_CHECK_XX
* Refreshed patch 0001-makedumpfile-replace-hardcode-CFLAGS.patch
* Refreshed patch 0002-mem_section-Support-only-46-bit-for-MAX_PHYSMEM_BITS.patch;
had conflict in patch removed function, is this patch still required?
* Removed 0001-PATCH-Remove-duplicated-variable-definitions.patch backported
patch
* Added dependency to xz required by new version
Signed-off-by: Alexander Kamensky <alexander.kamensky42@gmail.com>
makedumpfile can be compiled with lzo and/or snappy dump file compression
support. Lzo and snappy offer faster compression time but slightly bigger
compressed files. When reload time is important it might be beneficial to use
lzo or snappy compression.
But it brings additional dependencies and can make crash image bigger.
Add makedumpfile PACKAGECONFIG that would allow to add lzo and/or snappy if so
desired.
Signed-off-by: Alexander Kamensky <alexander.kamensky42@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Include lib and kernel subdirs as vm tests need headers from it in 5.8+
(test_hmm_uapi.h) see [1]
It now uses tests which are not compatible on arm starting with 5.8+
[1] https://lkml.org/lkml/2020/4/22/1654
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Sakib Sajal
Wang Mingyu
Randy MacLeod
Chase Qi
Khem Raj
Mingli Yu
Claus Stovgaard
William A. Kennington III
Fathi Boudra
Richard Purdie
Oleksandr Kravchuk
Matteo Croce
zangrc
Peter Morrow
Martin Jansa
Olivier Georget
Daniel Klauer
LiweiSong
Bruce Mitchell
INC@Cisco)
zhengruoqin
Beniamin Sandu
Alexander Kamensky