Accommodates the following warning issued by CMake v3.27.1.
CMake Deprecation Warning:
Compatibility with CMake < 3.5 will be removed from a
future version of CMake.
The underlying driver for this: Debian/Ubuntu's libbladerf1
package installs the udev rules itself, instead of having a
separate libbladerf-udev package like our PPA does. So,
having libbladerf1 installed creates a conflict with installing
libbladerf2, which is not good.
End result is that we either need to conflict libbladerf1, or
put the udev rules from libbladerf-udev somewhere else, and
there are fewer side effects with the latter option.
Builds were failing in various ways on 32-bit platforms, due to various assumptions (and the fact that pretty much everything is 64-bit nowadays). This PR should fix all of them so that it builds clean.
* host: libbladeRF/thirdparty/adi: fix build error on 32-bit systems
sizeof(void*) is not necessarily equal to sizeof(uint64_t)
* host: libbladeRF/doc: fix 32-bit build failure on doc example
sizeof(size_t) does not always equal sizeof(uint64_t)
* host: docker: remove centos-7-i386 build
it doesn't work, and this is an edgy-enough case that i think we'll be
okay with the ubuntu i386 builds
* host: libbladeRF/bladerf2: explicitly note constant types
On non-64-bit platforms, these integer constants can be ambiguous, which
throws compiler warnings.
* host: libbladeRF_test: explicitly note constant types
On non-64-bit platforms, these integer constants can be ambiguous, which
throws compiler warnings.
* host: docker: clean up the base build process a bit
* host: docker: reduce build context size
We don't need the entire tree (since the _base image has that), so we
can just build from the same dir as the Dockerfile and not transfer
everything to the daemon over and over again.
* host: docker: WIP: script for automated Dockerized testing
* host: docker: use git clean when building base image
* host: docker: add script to automatically freshen base images
* host: docker: auto-detect build dir when testing
* host: docker: build arg for make -j parameter
* host: docker: refactor build script
* host: docker: add parallel run script
* host: docker: adjust make -j settings for builds
* host: docker: test: WIP commit
* host: docker: add usb product check, rfic loopback only for micros
* host: docker: copy the FPGA images from the base
- new build distinguishers:
- compiler: gcc or clang
- buildtype: Release or Debug
- taggedrelease: NO or YES
- add com.nuand.ci.* labels for image metadata
- create a base image with a pruned checkout, to reduce disk consumption
and to snapshot the working directory's state at the start of a build
spree
- add -j$(nproc) to the make invocation
- reorganize Dockerfiles to improve build cache hits
- build.bash: iterate over rel, compiler, and buildtype, for a
combination of 6 builds per distro version
- build.bash: general restructuring for DRY
* host: misc: add Dockerfiles for multi-distro test builds
It is now possible to perform test builds automatically across various
releases of Arch, CentOS, Debian, Fedora, and Ubuntu with:
bash host/mis/docker/build.bash
* host: docker: fix ld.so.conf for centos-7
With the addition of a warning to the lms.c code, we've been seeing
some codes hit a max VCOCAP value before hitting a VCO limit, indicating
that we're too close to the edge of a band.
Some of the changes in the table entry occur close to the end of
a VCO's freq range (per the plot in the LMS6002D FAQ, section 5.24).
As such, it was observed that the PLLs could lose lock over temperature
fluctuations.
The table entries have been updated and made more conservative with
respect where the VCO selection and/or divider ("FREQRANGE") change is
made.
Additionally, the table format has been changed to make it clear which
VCO and associated divider is being used.
A big thanks goes out to the Yate.ro team for their efforts in
identifying and resolving PLL instability issues.
Most of these files had inaccurate copy-paste headers, as the files are
now vastly different than those mentioned. These will all just fall under
the "meta-data" category in upcoming licenses.
Per discussion with FindLibBladeRF.cmake author, the text in
FindLibTecla.cmake was unnecessary, and has been removed.
- Added method to libbladeRF
- Added new VID/PID pair to udev rule to prevent requiring root access
- Added jump_to_boot bladeRF-cli command
- Fixed most warnings in FX3 firmware