* CMakeLists.txt:
always pass "EXACT" to find_package(Python3).
because per cmake document, "EXACT" only takes effect when
<Package>_FIND_VERSION_COUNT is greater than 1, where <Package>
is "Python3". see also cmake/modules/FindPython/Support.cmake
* cmake/modules/AddCephTest.cmake:
drop redundant find_package(Python3) calls. since Python3 is
a mandatory requirement for building Ceph, we only need a
single call of find_package(Python3..) in the top of the source
tree. the only possible case to repeat it is to ensure that we
have the correct version of Python3 used in following CMake
script. but there is no need to repeat it if we just want to
ensure that we have a python3 interpretor in place.
* cmake/modules/Distutils.cmake:
always pass "EXACT" to find_package(Python3).
we should always pass EXACT to find_package() when finding python3,
this is a follow-up of e2babdfae8
Signed-off-by: Kefu Chai <tchaikov@gmail.com>
The purpose of this patch is to add the initial support to
offload memory/pmem operations by sync usage through hardware path
in DML library.
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
if another python3 with higher version is found by
find_package(Python3), the cmake's install script would just
install the python modules/extensions into that python3's
dist-package directory, and the packaging script would fail
to find these artifacts when trying to package them.
so we need to ensure that the install directories for python
modeules/extensions are always "versioned" with WITH_PYTHON3
cmake option.
Signed-off-by: Kefu Chai <tchaikov@gmail.com>
to address following warning from python 3.9:
<string>:1: DeprecationWarning: The distutils package is deprecated and slated for removal in Python 3.12. Use setuptools or check PEP 632 for potential alternatives
<string>:1: DeprecationWarning: The distutils.sysconfig module is deprecated, use sysconfig instead
Signed-off-by: Kefu Chai <tchaikov@gmail.com>
adds utf8proc submodule, needed by the arrow submodule in centos. add a
WITH_SYSTEM_UTF8PROC option that controls whether or not utf8proc is
built from submodule
non-system utf8proc is built as a static library to avoid conflicts with
system-provided libraries
ceph.spec.in sets WITH_SYSTEM_UTF8PROC=OFF until it's available in
centos
Signed-off-by: Casey Bodley <cbodley@redhat.com>
adds an arrow submodule. when WITH_RADOSGW_SELECT_PARQUET is enabled,
the submodule is built as an external project and rgw links against its
imported Arrow::Parquet target
Signed-off-by: Casey Bodley <cbodley@redhat.com>
ceph is still using the Makefile based building system for building
DPDK. and DPDK enables -Werror if RTE_DEVEL_BUILD is 'y' which is
enabled by default when the dpdk is built from a git repo.
but newer GCC is more picky than the older versions, to prevent
the possible FTBFS when we switch to newer GCC for building old
branches whose dpdk submodule might be include the changes addressing
those warnings. let's just disable this option.
the only effect of this option is to add -Werror to CFLAGS. but
the building warnings from DPDK is not our focus when developing
Ceph in the most cases. so it should be fine.
see also
eac901ce29/doc/build-sdk-quick.txt (L18)
Signed-off-by: Kefu Chai <tchaikov@gmail.com>
* since the minimum required cmake version is now 3.16, there is no
need to check for cmake version for using multiple URLs for
downloading external project.
* use string(JOIN ..) to compose URL option for downloading boost
Signed-off-by: Kefu Chai <tchaikov@gmail.com>
because QAT driver with version v1.7.l.4.14.0 or higher cannot be dowmloaded
directly by URL, FindQatDrv.cmake can find the locally installed QAT package and libraries
Signed-off-by: Miaomiao Liu <miaomiao.liu@intel.com>
Signed-off-by: Hualong Feng <hualong.feng@intel.com>
We build spdk as static library, linking against them requires the
use of `-Wl,--whole-archive` as argument, otherwise we will have error
`nvme.c: nvme_probe_internal: *ERROR*: NVMe trtype 256 not available`.
This is due to the use of constructor functions in spdk to register
NVMe transports. So we need to do so to ensure we call all the
constructors.
Signed-off-by: Tongliang Deng <dengtongliang@sensetime.com>
* add find_package() support for detecting the existence of
linux/blkzoned.h before using it.
* link against Linux::ZNS for adding the compilation definition of
HAVE_ZNS instead of including it in the global config.h
* move the CMake option closer to where it is used for
better readability. as this option takes effect only if
crimson is compiled.
* make WITH_ZNS an option which depends on the value of
CMAKE_SYSTEM_NAME. it is hidden and off if CMAKE_SYSTEM_NAME
is not "Linux".
Signed-off-by: Kefu Chai <tchaikov@gmail.com>
migrate from using opentracing-cpp to opentelemetry-cpp static as distributed tracing API
Reviewed-by: Yuval Lifshitz <ylifshit@redhat.com>
Reviewed-by: Josh Durgin <jdurgin@redhat.com>
replaced by Opentelemetry project, remove building and linking of jaeger
libraries.
* remove externalProject building
* linking of jaegertracing dependencies to ceph project
Signed-off-by: Deepika Upadhyay <dupadhya@redhat.com>
we no longer need to package libjaeger, as we would be using
libopentelemetry static libraries(until we support shared).
Signed-off-by: Deepika Upadhyay <dupadhya@redhat.com>
we build and use targets needed for jaeger to work with opentelemetry
referring:
https://github.com/open-telemetry/opentelemetry-cpp/blob/main/examples/jaeger/CMakeLists.txt
right now we need a patch for using thrift from distro packages, hence
we are using GIT_FETCH based externalProject method, for development
purpose, which will be replace by more reliable method in future.
NOTE: phase 1, still needs to be tested.
Signed-off-by: Deepika Upadhyay <dupadhya@redhat.com>
it's reported that a mips64el build host is able to pass the test of
CheckCxxAtomic without linking against libatomic, while librbd.so
fails to link due to failures like
/usr/bin/ld: ../../../lib/librbd.so.1.16.0: undefined reference to `__atomic_store_16'
/usr/bin/ld: ../../../lib/librbd.so.1.16.0: undefined reference to `__atomic_load_16'
/usr/bin/ld: ../../../lib/librbd.so.1.16.0: undefined reference to `__atomic_compare_exchange_16'
so we have to check the existence of __atomic_load_16 instruction on
mips architecture.
Signed-off-by: Kefu Chai <tchaikov@gmail.com>
now that python3.10 has two digits as its minor version, we should be
able to match it.
this change fixes the build with python3.10. without this change, we'd
have
error: wrong library name 'python310' in the --with-<library> option.
when building with python3.10
Signed-off-by: Kefu Chai <tchaikov@gmail.com>
Whereas new LevelDB packages are build with -fno-rtti and break our
attempts to compile against them,
and whereas LevelDB has been deprecated for some time, with the Ceph
team expending great effort to migrating OSDs off of it,
let it be therefore removed.
Signed-off-by: Adam C. Emerson <aemerson@redhat.com>
The common_async_dpdk library depends on the dpdk librte_eal file.
If the static library is used, the dependency is transferred to
libceph-common. By default, libceph-common uses the PUBLIC keyword,
the keyword PUBLIC determines that both the application and
libceph-common link to dpdk librte_eal.a. As a result, the global
variables in dpdk librte_eal.a has two copies. So link DPDK shared
libs to avoid DPDK initialization failure.
Fixes: https://tracker.ceph.com/issues/42861
Signed-off-by: Chunsong Feng <fengchunsong@huawei.com>
Signed-off-by: luorixin <luorixin@huawei.com>
* dbstore also has tracing dependencies and missing linking for libraries
* use `using` keyword in common/tracer.cc to silence unused method
warnings
* log building of external project jaegertracing, opentracing into build
file, (done to reduce external project build verbosity)
Signed-off-by: Deepika Upadhyay <dupadhya@redhat.com>
* cmake/modules/SIMDExt.cmake: define HAVE_PPC for 32-bit PowerPC.
* src/arch/CMakeLists.txt: compile ppc.c for all PowerPC architectures,
including powerpc (32-bit PowerPC), ppc64el (64-bit Little Endian
PowerPC) and ppc64 (64-bit Big Endian PowerPC).
before this change, ppc.c is only compiled if HAVE_POWER8 is defined.
but Power8 is a 64-bit PowerPC architecture. while in src/arch/probe.cc,
we check for `defined(__powerpc__) || defined(__ppc__)`, if this is
true, ceph_arch_ppc_probe() is used to check for the support of
Altivec. but on non-power8 PowerPC machines, the linker fails to find the
symbols like ceph_arch_ppc_probe(), as ppc.c is not compiled on them.
in this change, ppc.c is compiled on all PowerPC architectures, so that
ceph_arch_ppc_probe() is also available on non-power8 machines. this
change does not impact the behavior of non-power8 machines. because
on them, the runtime check would fail to detect the existence of
PPC_FEATURE2_VEC_CRYPTO instructions.
Reported-by: Mattias Ellert <mattias.ellert@physics.uu.se>
Signed-off-by: Kefu Chai <tchaikov@gmail.com>
to address following warning:
fatal: reference is not a tree: 7f02f21f53ccd5e2448086f8e9015489693dd2dc
CMake Error at /home/jenkins-build/build/workspace/ceph-pull-requests/build/fio_ext-prefix/tmp/fio_ext-gitclone.cmake:40 (message):
Failed to checkout tag: '7f02f21f53ccd5e2448086f8e9015489693dd2dc'
it seems that the shallow option does not work with a sha1 tag option,
let's continue using the ceph repo with a tag.
Signed-off-by: Kefu Chai <tchaikov@gmail.com>
this change partially reverts 10baab3fc8,
but since the fix for C++ build is not included by any tag or branche so
far. let's just use the sha1 for now.
Signed-off-by: Kefu Chai <tchaikov@gmail.com>
CMAKE_THREAD_LIBS_INIT can be empty. On newer platforms, this leads to
an error:
CMake Error at cmake/modules/Buildpmem.cmake:47 (set_target_properties):
set_target_properties called with incorrect number of arguments.
This is because glibc 2.34 merged libpthread.so into libc.so.
Fedora 35 and RHEL 9 have glibc 2.34.
Fixes: https://tracker.ceph.com/issues/52353
Signed-off-by: Ken Dreyer <kdreyer@redhat.com>
Co-authored-by: Kefu Chai <kchai@redhat.com>
y2c.py is like a compiler which translates .yaml to .cc and .h files,
it does not have access to all .yaml files. to validate the dangling
see-also issue, we need to do this with a "linker".
in this change, validate-options.py is introduced to check if any of
option name included by the see-also property is valid.
Fixes: https://tracker.ceph.com/issues/51483
Signed-off-by: Kefu Chai <kchai@redhat.com>
when download/building grafonnet-lib, dpdk, spdk, liburing and fio,
they dump lots of output during configuration and building phrases,
all of which is irrelevant to us. so let's just silence it.
Signed-off-by: Kefu Chai <kchai@redhat.com>
to silence warnings like:
In file included from nvme_tcp.c:52:
/home/jenkins-build/build/workspace/ceph-pull-requests/src/spdk/include/spdk_internal/nvme_tcp.h:369:17: warning: variable 'plen' set but not used [-Wunused-but-set-variable]
uint32_t hlen, plen;
^
Signed-off-by: Kefu Chai <kchai@redhat.com>
otherwise we have:
src/spdk/dpdk/lib/librte_eal/linux/eal.c:591:11: error: variable 'total_mem' set but not used [-Werror,-Wunused-but-set-variable]
uint64_t total_mem = 0;
^
1 error generated.
when building dpdk with Clang. dpdk passes -Werror to compiler, so
we have to disable this option.
Signed-off-by: Kefu Chai <kchai@redhat.com>
to pick up the clang build fix. we should use the upstream repo, once the
clang build fix gets merged.
bumping up the fio helps to address following error, as this part was rewritten:
src/test/fio/CMakeFiles/fio_ceph_objectstore.dir/fio_ceph_objectstore.cc.o.d -o src/test/fio/CMakeFiles/fio_ceph_objectstore.dir/fio_ceph_objectstore.cc.o -c ../src/test/fio/fio_ceph_objectstore.cc
In file included from ../src/test/fio/fio_ceph_objectstore.cc:26:
In file included from src/fio/fio.h:18:
In file included from src/fio/thread_options.h:6:
In file included from src/fio/options.h:8:
src/fio/parse.h:128:13: error: arithmetic on a pointer to void
return ret + offset;
~~~ ^
1 error generated.
Signed-off-by: Kefu Chai <kchai@redhat.com>
on Clang, the option for detecting unknown attribute is
-Wunknown-attributes, so "-Wattributes -Werror" does not fail the test
when the C compiler is Clang.
in this change, we just turn all warnings into errors.
this should fail the test if the compiler does not understand
`__attribute__((__symver__ ...))`
Fixes: https://tracker.ceph.com/issues/40060
Signed-off-by: Kefu Chai <kchai@redhat.com>
We should check if -flto-partition=none is defined when the compiler
does not support symver attribute and fail the build if it is not.
Fixes: https://tracker.ceph.com/issues/40060
Co-authored-by: Kefu Chai <tchaikov@gmail.com>
Signed-off-by: Boris Ranto <branto@redhat.com>
