Re: Error cross-compiling OOT modules for E310

[Ivan Iudice]

Anyone who is developing on E310?

Ivan

> Il giorno 12 feb 2020, alle ore 17:40, kron...@tiscali.it ha scritto:
> 
> 
> Hi guys,
> I'm trying to compile one of mine OOT module (C++), but I'm obtaining a 
> curious error.
> https://pastebin.com/ZyfJcHCa
> Looking at warnings, it seems that several header files are got from 
> /usr/include, and not from sdk sysroot.
> What's going on?
> Thank you.
> Ivan
> 
> 
> 
> Con Tiscali Mobile Smart 30 4G hai minuti illimitati, 100 SMS e 30 Giga in 4G 
> a soli 8,99€ al mese. http://tisca.li/smart30
> 

Re: UHD and USRP1

[Michael Margaras]

 Hello Marcus, Hello Geof,

Thanks a lot for your replies !

Indeed tried it on a Ubuntu 19.10 and everything looks fine there with latest 
UHDs.
Benchmark_rate does not report any dropped samples even at 8 MSPS . 

On W10, indeed I also got very nice audio results (no clicks, interruptions 
etc) while running an FM receiver example with USRP1 streaming at 4MSPS.
Up to that time I thought it's just a reporting issue.
But when I tried some NOAA HRPT satellite reception where rates are higher, it 
did't work well at all indicating that this might be a real issue.

Good to know that on Linux it works! That's a relief ! :-)

Thanks,
Michael Στις Δευτέρα, 17 Φεβρουαρίου 2020, 12:01:52 π.μ. EET, ο χρήστης 
Geof Nieboer  έγραψε:  
 
 Michael,
I still have my USRP1, and ran the same test w/ 3.14 and got the same results 
on Win10, dropped packets down to the minimum supported rate.  Same test, same 
driver against the B200 works fine up to about 20MSps which appears to be my 
USB 3 chip's limit.
For kicks I downloaded the same older version of UHD you had success with, and 
ran it without a problem up to 8 MSps.

For "real" usage, the issue is similar... I just checked that the UHD_WBFM 
sample .grc file still works with the latest UHD drivers, etc, audio is fine 
(even if I kick up the samp_rate to 4MSps) though the same "D" dropped packets 
are reported.
My VirtualBox installation needs to be updated, or I'd check in Linux as well.
So I think it's a driver issue, though if it's really dropping packets or just 
reporting them dropped is a question for Ettus.
Geof
On Sun, Feb 16, 2020 at 11:52 AM Marcus Müller  wrote:

Hi Michael,

GR 3.3?! Wow, that *is* retro :)

> Now I wanted to give it a try on W10 and a modern PC, but when I run a
RX benchmark_rate test with latest UHD 3.15 but also 3.10  the command
window gets filled up with DD messages and number of dropped samples
is huge.

That shouldn't happen. The maximum rates of the USRP1 should never even
stress a modern PC.

BTW, not even sure libusrp1 had indications for dropped samples. In
fact, I remember the documentation back in the day of GR 3.4's gr-usrp
said it would simply drop samples on multi-device reception until
streams aligned. Don't remember whether that included any signalling.

Anyway, I suspect USB strangeness under windows 10.

Did you try UHD 3.15 on Linux? If you just set up a Ubuntu 19.10 system,
the GNU Radio you can install directly through apt-get there will bring
in UHD 3.15.
I don't have a USRP1 at hand, so I can't test this :(

Best regards,
Marcus

On 16.02.20 07:51, Michael Margaras wrote:
> Hello All,
> I have been using my USRP1 mainly on an older Linux PC with GNU radio 3.3 .
> Now I wanted to give it a try on W10 and a modern PC, but when I run a RX 
> benchmark_rate test with latest UHD 3.15 but also 3.10  the command window 
> gets filled up with DD messages and number of dropped samples is huge.
> 
> Have also tried the older UHD 3.005.004 on the same PC and there I am no 
> getting any dropped samples !
> 
> Anyway , with UHD 3.15 an 3.10, at the end of the test , the number of 
> dropped samples is huge but remarkably overflows are 0.
> This happens at all sample rates but the number of dropped samples  is 
> decreasing as the sample rates decreases but still remains huge even a lowest 
> sampling rate.
> 
> On the other hand , on an older Dell D630 PC running  Ubuntu 9.10 and GNU 
> Radio 3.3 (libusrp) no dropped samples or overflows have been observed up to 
> almost 8 MSPS.
> 
> Please find benchmark_rate results of the W10 PC using UHD 3.10 at the end of 
> this message for 6MSPS , 2MSPS and 0.1875 MSPS.
> BTW, master clock was modified to 48 MHz but same results have been observed 
> with the default 64 MHz clock.
> 
> Is anybody successfully using USRP1 on W10 or Linux with latest UHD releases? 
> How can this dropped-samples issue be fixed?
> 
> Any suggestion would be very welcome.
> 
> Many thanks in advance !
> 
> Michael Margaras
> SV1CAL
> 
> 
> Below testing 6 MSPS , 2 MSPS  and 0.1875 MSPS :
> 
> 
> C:\Program Files\UHD\lib\uhd\examples>benchmark_rate --rx_rate 6e6 --rx_cpu 
> sc16 --rx_otw sc16
> Win32; Microsoft Visual C++ version 14.0; Boost_106000; 
> UHD_003.010.001.001-0-unknown
> 
> 
> -- Loading firmware image: C:\Program 
> Files\UHD\share\uhd\images\usrp1_fw.ihx... done
> *** Warning! ***
> Benchmark results will be inaccurate on USRP1 due to insufficient features.
> 
> Creating the usrp device with: ...
> -- Opening a USRP1 device...
> -- Loading FPGA image: C:\Program 
> Files\UHD\share\uhd\images\usrp1_fpga.rbf... done
> -- Using FPGA clock rate of 48.00MHz...
> Using Device: Single USRP:
>   Device: USRP1 Device
>   Mboard 0: USRP1
>   RX Channel: 0
>     RX DSP: 0
>     RX Dboard: B
>     RX Subdev: TVRX
>   TX Channel: 0
>     TX DSP: 0
>     TX Dboard: A
>     TX Subdev: WBXv2 TX+GDB
> 
> Setting device timestamp to 0...
> Testing receive rate 6.00 Msps on 1 channels
> D

Is this the right way to a frequency estimator?

[sarandis. Doulgeris]

[GNU Radio 3.8] Error loading modules on E310

[krono86]

  
Dear all, 
finally I cross-compiled my OOT module for running on
USRP E310. 
Based on the instructions at
"https://kb.ettus.com/Software_Development_on_the_E3xx_USRP_-_Building_RFNoC_UHD_/_GNU_Radio_/_gr-ettus_from_Source";,
I set the environment variable PYTHONPATH a little bit different, to
point the path of the installed OOT module: 

export
PYTHONPATH=$LOCALPREFIX/lib/python3.5/dist-packages:$PYTHONPATH

Such a
way, on the target I can load my custom module in python.

root@ni-e31x-316AFEA:~# python3 -c "import custom_mod" 
I created a
flowgraph that use my OOT module, and I discovered in the newer file
system for E310 python3 has not all of the needed libraries.

root@ni-e31x-316AFEA:~# ./top_block.py
Traceback (most recent call
last):
 File "./top_block.py", line 12, in 
 from gnuradio import
blocks
 File
"/home/root/localinstall/usr/lib/python3.5/dist-packages/gnuradio/blocks/__init__.py",
line 38, in 
 from .stream_to_vector_decimator import *
 File
"/home/root/localinstall/usr/lib/python3.5/dist-packages/gnuradio/blocks/stream_to_vector_decimator.py",
line 23, in 
 from gnuradio import gr
 File
"/home/root/localinstall/usr/lib/python3.5/dist-packages/gnuradio/gr/__init__.py",
line 46, in 
 from .top_block import *
 File
"/home/root/localinstall/usr/lib/python3.5/dist-packages/gnuradio/gr/top_block.py",
line 32, in 
 from .hier_block2 import hier_block2
 File
"/home/root/localinstall/usr/lib/python3.5/dist-packages/gnuradio/gr/hier_block2.py",
line 26, in 
 import pmt
 File
"/home/root/localinstall/usr/lib/python3.5/dist-packages/pmt/__init__.py",
line 61, in 
 from .pmt_to_python import pmt_to_python as to_python

File
"/home/root/localinstall/usr/lib/python3.5/dist-packages/pmt/pmt_to_python.py",
line 23, in 
 import numpy
ImportError: No module named 'numpy' 
How
could I solve the problem? 
I'm so close to run my OOT modules on the
target... 
Thanks so much. 
Ivan  


Con Tiscali Mobile Smart 30 4G hai minuti illimitati, 100 SMS e 30 Giga in 4G a 
soli 8,99€ al mese. http://tisca.li/smart30

multiple messages in GRC as different commands

[Ali G. Dezfuli]

Hi all,
I wonder if it is possible to command a "Signal Source" block in GRC to
change its frequency to several values with the messages.
For example, I want to change its frequency to 100 different values during
run-time using available blocks (if possible).
thanks in advance

Re: Aw: Re: AttributeError: OOT Modul

[Vasil Velichkov]

Hi Till,

Please always try to keep the mailing list in the loop and do not reply 
privately.

Here is how to demangle these undefined symbols

  $ c++filt _ZN3uhd4usrp10multi_usrp4makeERKNS_13device_addr_tE
  uhd::usrp::multi_usrp::make(uhd::device_addr_t const&)

  $ c++filt 
_ZN3uhd13device_addr_tC1ERKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEE
  uhd::device_addr_t::device_addr_t(std::__cxx11::basic_string, std::allocator > const&)

You need to link the uhd library to your OOT module. In your lib/CMakeLists.txt 
you need something like

  target_link_libraries(gnuradio-control
UHD::UHD
#other libraries
  )

Best Regards,
Vasil

On 17/02/2020 16.00, "Till Hülder" wrote:
> Hi Vasil,
> i do get any undefined symbols:
> iheamira@iheamira-MS-7885:~/Dokumente/BA_Huelder/GPIO_modul/gr-control/build$ 
> ldd -r lib/lib*.so
> lib/libgnuradio-control-1.0.0git.so:
> undefined symbol: _ZN3uhd4usrp10multi_usrp4makeERKNS_13device_addr_tE   
>   (lib/libgnuradio-control-1.0.0git.so)
> undefined symbol: 
> _ZN3uhd13device_addr_tC1ERKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEE
> (lib/libgnuradio-control-1.0.0git.so)
> undefined symbol: _ZN3uhd4usrp10multi_usrp4makeERKNS_13device_addr_tE   
>   (lib/libgnuradio-control.so)
> undefined symbol: 
> _ZN3uhd13device_addr_tC1ERKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEE
> (lib/libgnuradio-control.so)
> iheamira@iheamira-MS-7885:~/Dokumente/BA_Huelder/GPIO_modul/gr-control/build$ 
> ldd -r swig/*_swig.so
> undefined symbol: _ZN3uhd4usrp10multi_usrp4makeERKNS_13device_addr_tE   
>   (/usr/local/lib/libgnuradio-control-1.0.0git.so.0.0.0)
> undefined 
> symbol:_ZN3uhd13device_addr_tC1ERKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEE
>
>   (/usr/local/lib/libgnuradio-control-1.0.0git.so.0.0.0)
> How can i fix this?
> Best regards,
> Till
> *Gesendet:* Mittwoch, 12. Februar 2020 um 18:12 Uhr
> *Von:* "Vasil Velichkov" 
> *An:* "Till Hülder" 
> *Cc:* discuss-gnuradio@gnu.org
> *Betreff:* Re: AttributeError: OOT Modul
> Hi Till,
> 
> On 12/02/2020 16.46, "Till Hülder" wrote:
>  > Hello,
>  > i build a OOT Module and after installing i get this error when i run the 
> program .
>  > File "/home/iheamira/Dokumente/BA_Huelder/gnu-radio/top_block.py", line 
> 219,
>  > in 
>  > main()
>  > File "/home/iheamira/Dokumente/BA_Huelder/gnu-radio/top_block.py", line 
> 207,
>  > in main
>  > tb = top_block_cls()
>  > File "/home/iheamira/Dokumente/BA_Huelder/gnu-radio/top_block.py", line 
> 142,
>  > in __init__
>  > self.control_control_b_0 = control.control_b(5e-6, 1000, 10, 5e-6, 5e-6,
>  > 5e-6, 200e-6, 100e-6, 0.2, samp_rate)
>  > AttributeError: 'module' object has no attribute 'control_b'
> 
> One possible reason for such error is that there are undefined symbols in 
> your 
> shared library.
> 
> Go to the build directory and execute
> 
> ldd -r lib/lib*.so
> ldd -r swig/*_swig.so
> 
> and if you see "undefined symbol:" messages that means you are not linking 
> all 
> dependencies.
> 
> Also in `python/__init__.py` you probably have a try except block like this 
> one
> 
> # import swig generated symbols into the OOT namespace
> try:
> # this might fail if the module is python-only
> from .OOT_swig import *
> except ImportError:
> pass
> 
> Consider removing the try: and except and keep only the import line
> 
> from .OOT_swig import *
> 
> Read also:
> https://lists.gnu.org/archive/html/discuss-gnuradio/2020-02/msg00047.html
> https://lists.gnu.org/archive/html/discuss-gnuradio/2020-02/msg00054.html
> 
> Regards,
> Vasil
> 

Re: About the transceiver and sending packets to / from a hardware

[Bastian Bloessl]

Hi,

you're right, the RIME stack is a network stack. But it is completely
optional. You can just disable the block if you don't need it. It was
just convenient to setup connections with Contiki running on a TelosB node.

If you use something different than Contiki (i.e., a node w/o RIME)
frames will not pop out of the RIME ports. But you will see all frames
that are received by the SDR in the Wireshark trace. It logs everything
that is received, independent from PAN ID, etc.

If you use other nodes, as Marcus said:
- check that they use the same PHY (O-QPSK in 2.4GHz)
- check that they use the same channel
- To RX with a custom node: implement a firmware that shows all frames
(independent from MAC, PAN ID, etc.)
- To RX frames w/ GNU Radio, you don't have to do anything special as
every received frame will show up in GR.

Some nodes with PCB antennas are not that great. If you have problems
receiving them with GR, try to (1) increase TX gain, (2) get it closer
to the SDR, (3) increase RX gain of the SDR.

Hope it helps,
Basti

On 2/15/20 9:51 PM, sampath ranga wrote:
> Hello Marcus Muller , 
> 
>    Thank you so much for the reply . The thing about Rime Stack is rime
> stack is a network level stack . So my question is if its network level
> stack is that the block thats restricting me from getting other hardware
> signals on the Wireshark ?? and I want to let you know that i am trying
> different type of flow graph changes to make it work and i am posting
> this after playing with MAC and Rime . I even tried to seperate the tx
> and rx section to see if it is working but still it is not . 
>  
> 
> Thanks and regards ,
> Ranganathan Sampathkumar 
> 
> On Sat, Feb 15, 2020 at 2:39 PM Marcus Müller  > wrote:
> 
> Hi,
> 
> slightly will have to repeat myself:
> 
> On 15.02.20 21:15, sampath ranga wrote:
> > Hello Everyone,
> >
> >     1. I was trying for the past few weeks to make the Zigbee Trx
> > communicate with the other Zigbee hardware for implementation
> > verification . I am not using TelOS b mote with COntiki Firmware
> > instead i am using a Zigbee dongle as a hardware to make the
> > communication work . My question is why is the flowgraph restricting
> > only to hear the "hello world" message but not the message from other
> > hardware? If my SDR could operate at 2.4GHz in air , the transceiver
> > block should not only hear the "GNU Radio Transmitted message but also
> > from the other hardware isnt it? The other hardware is also not
> > receiving the GNU Radio's message.
> 
> This is wireless communications. Not only do you have to configure all
> devices in a compatible manner, channel and receiver conditions can make
> interoperation problematic.
> 
> You will have to apply your own signal processing / wireless education
> to debug this.
> 
> >   2. Is the Rime stack / MAC layer that fixes frames stopping the
> > other hardware to send or receive?
> 
> 
> I don't understand this question, sorry.
> 
> >   3. I am also trying the same Zigbee Trx in 2 laptops, the message
> > that i set in laptop 1 at GNU Radio Zigbee TRX has been received in
> > wireshark at laptop 1 and laptop 2 message at laptop 2 , Why are not
> > they syncing with each other if i give same destination or pan id with
> > channel number?
> 
> Same answer as to 1. You're not doing any signal-level debugging on
> this, so we can't help you.
> 
> > Technically i am trying to do 2 work , Either have two laptop that has
> > each flowgraph with 2 different message and wanna see if they both are
> > transferring their message at wireshark, which is right now not
> happening
> Same answer as to 1.
> > secondly , trying to transmit Zigbee Trx basti's message to a seperate
> > zigbee hardware (Xbee or any zigbee sniffer) or send message from Xbee
> > and trying to receiving in on Wireshark at the GNU Radio end
> 
> Same answer as to 1.
> 
> > I also face the issue of when i transceive the message its showing
> > "Bad FCS"  and also the packet flow is getting stopped when i open the
> > wireshark . Please throw some ideas on this . 
> >
> Same answer as to 1.
> 

-- 
Dr. Bastian Bloessl
Secure Mobile Networking Lab (SEEMOO)
TU Darmstadt, Germany

www.bastibl.net
GitHub/Twitter: @bastibl

Re: Kurtosis estimator

[Martin Braun]

There's also kurtotic_equalizer_cc, if that's what you're looking for.

On Mon, Feb 10, 2020 at 2:11 AM Müller, Marcus (CEL) 
wrote:

> Maybe you can abuse the mpsk_snr_est for your purposes?
>
> On Sun, 2020-02-09 at 22:16 -0500, Marcus D. Leech wrote:
> > Has anyone implemented a Kurtosis estimator in GR?  Maybe a convenient
> OOT?
> >
> >
> >
>

Project Call this Thursday, 10 AM Pacific, 19:00 CET

[Martin Braun]

Hi all,

our monthly project call is this Thursday, 10 AM Pacific Time, 19:00 CET, 1
PM Eastern Time.
Please head over to IRC or Slack for the coordinates. Like last time, we
intend to stream the call on Twitch, but you can still join the call
directly. Note we're still getting used to Twitch chat as yet another chat
medium, maybe eventually we'll figure out how to sync chats. Bear with us
as we catch up with 2020 technology (twitch.tv/gnuradio).

Cheers,
Martin

Using ZMQ source blocks with a separate process

[Barry Duggan]

Hi,

I would like to use one of the ZMQ source blocks (such as ZMQ PULL 
Message Source) to communicate with a separate Python process. I have 
been able to use various Python examples to test between processes, but 
when I try to use a GRC block, it fails. Are there restrictions which 
prohibit data exchange outside the gnuradio realm?


Thanks for your help.
--
Barry Duggan KV4FV