[UHD 3.8.0] Release Announcement

Hello lists,

at this point, we are releasing our UHD version 3.8.0. There were some
minor changes since the release candidate, but mainly regarding
documentation, the Boost version and minor build fixes to make sure UHD
works on all platforms.

You can get this release by either pulling master branch, or checking
out the release’s tag:
https://github.com/EttusResearch/uhd/tree/release-003_008_000

If you haven’t pulled master for a while, it will be a huge diff (might
take a bit, don’t worry). A couple of notes on this release:

  • We have reorganized the firmware and fpga subdirectories. The former
    was simply moved around, but the latter is now a git submodule. Run ‘git
    submodule init’ and ‘git submodule update’ if you want to read and
    change the fppga source code. As a result, the source tarballs will be
    much smaller.
  • B200 updates: We’ve fixed the issue with B200 rapidly losing
    connection on some devices (this particular issue never existed on
    maint, but for a while it was in master). The reason was the version of
    the SDK we used to compile the firmware. Altogether, B200 initialization
    is now much faster.
  • E310: This release includes support for our upcoming embedded device.
    News on the actual device will follow in the near future, so please be
    patient.
  • CMake improvements: Among some fixes, we now distribute a
    UHDConfig.cmake file which you can use to more easily build applications
    linking to UHD. We provide an example on how to do that in
    host/examples/init_usrp/.

We will be resetting maint branch to master soon. From then on, the
development will happen as usual:

  • All bug fixes go into maint. This branch can be considered ‘stable’.
  • New features etc. go into master. This branch should be considered
    ‘unstable’.

Binary installers will follow this week.
More features and products are coming up!

Cheers,
Martin

Changelog for 3.8.0

  • Added E310 support
  • B200/B210: Moved AD9361 controls from firmware to host
  • Added several tools: ZPU dissector, improved CHDR dissector,
    kitchen sink, B200/B210 USB debugging utility, latency
    measurement tool.
  • Reorganized firmware/ directory structure. Refactored some
    firmware.
  • Removed FPGA sources, is now in own repository (submoduled).
  • Cleaned up command line arguments for some tools
  • Added math namespace, plus a unified float comparison infrastructure
  • Fixed tuning-related bugs
  • Moved manual over to Doxygen, also several manual bug fixes and
    amendments
  • Added many missing virtual destructors (less build warnings)
  • Added support for NI-RIO 14.0
  • X300 fixes: Not found over PCIe with no eth interfaces
  • CMake improvements: Now comes with own UHDConfig.cmake and example
    to build standalone UHD apps, build fixes on Apple, interoperability
    with GNU Radio
  • OctoClock fixes and improvements: Ethernet initialization, external
    ref detection, stability fixes, host driver (UHD can now talk to
    OctoClock)
  • Examples: Improved GPIO example, rx_samples_to_file
  • Bumped minimum Boost version to 1.46

On 10/28/2014 07:58 AM, Martin B. wrote:

If you haven’t pulled master for a while, it will be a huge diff (might
take a bit, don’t worry). A couple of notes on this release:

  • We have reorganized the firmware and fpga subdirectories. The former
    was simply moved around, but the latter is now a git submodule. Run ‘git
    submodule init’ and ‘git submodule update’ if you want to read and
    change the fppga source code. As a result, the source tarballs will be
    much smaller.

I’m trying to update gentoo but I don’t actually see an update for the
firmware images here:

http://files.ettus.com/binaries/master_images/

I only see rc1. Also of note, typically releases are always put into
archive/ at the same time so we have a link that will be valid for the
longer term.

Please advise.

Thanks,
Zero_Chaos

On 10/29/2014 03:35 PM, Rick F. wrote:

I only see rc1. Also of note, typically releases are always put into
archive/ at the same time so we have a link that will be valid for the
longer term.

Thanks for pointing this out – this was another change. First, for
anyone using uhd_images_downloader, nothing changes. That’ll still work.
Also, I have added the files to the usual location, so the manual
download folks will not have to sift through a long list of images.
In fact, for 99% of the UHD users population, nothing changes.

So, what did change? Mainly, our main place for images is >
files.ettus.com:/binaries/images/ from now on; files will not get
moved to an archive later on.

A quick motivation (and remember this is for a few people out there who
want to package stuff etc.): We want to be able to deploy a valid image
set for any given commit, and any branch – not just maint and master.
This is of course for our dev process, but we also may be releasing
other branches to the public in the future.
In the directory images/ in the repo, you’ll find a set of tools to grab
the correct image set for any given commit (starting with 3.8). Running
populate_imgs.py will dump a valid set of images into your soure tree.
We internally use this to generate the publically available .debs and
.rpms.

I hope this answers this question!

Cheers,
Martin

On 10/29/2014 01:10 PM, Martin B. wrote:

files.ettus.com:/binaries/master_images/

So, what did change? Mainly, our main place for images is >
files.ettus.com:/binaries/images/ from now on; files will not get
moved to an archive later on.

:frowning:

https://bugs.gentoo.org/show_bug.cgi?id=527650

Earlier this week there was a tar.gz with the following properties:

uhd-images_003.008.000-release.tar.gz 24794288 SHA256
d20c6105010a8fad66cadbeac38e99d0a63ffb7db8fa1b0161e4996df02416a2 SHA512
735577f9bf62edbdfd3b6a6f01cc8b15d01efe6f74ab9330481db478e3763118d269f7e67f02a22c02a5ab967f0e207541f663ef64279e0291654e94e0d40d17
WHIRLPOOL
2edac47f485dca697f2d125e1556fb161e0f0edbbd8827494dcc5764e968ae8c2d67316b065059e84912eecf7056da4bc5bd612939b6a987fb96909e4dfa821e

Now the tar.gz has been removed, I’m hoping this was an accident. Can
you restore the file? I have the original if you need it. I very much
dislike seeing these kinds of bugs :frowning:

-Zero_Chaos