I reproduced the same issue (except I completely uninstalled GR, UHD,
gr-ettus, etc. before doing rm -rf rfnoc then installed rfnoc from scratch)
then checked out 'eb5eed' in uhd-fpga which fixed the FPGA compatibility
number issue but there seems to be an underlying problem with that. I can't
use the RFNoC Radio block in receive mode. For example, I ran
rfnoc/src/gr-ettus/examples/rfnoc/rfnoc_ddc.grc with everything unmodified,
out-of-the-box and got this error:
[WARNING] [X300 RADIO] set_rx_gain: could not apply gain for this
daughterboard.
Traceback (most recent call last):
File "/home/switchlanez/rfnoc/src/gr-ettus/examples/rfnoc/rfnoc_ddc.py",
line 301, in <module>
main()
File "/home/switchlanez/rfnoc/src/gr-ettus/examples/rfnoc/rfnoc_ddc.py",
line 289, in main
tb = top_block_cls()
File "/home/switchlanez/rfnoc/src/gr-ettus/examples/rfnoc/rfnoc_ddc.py",
line 102, in __init__
self.uhd_rfnoc_streamer_radio_0.set_rx_antenna("TX/RX", 0)
File
"/home/switchlanez/rfnoc/lib/python2.7/dist-packages/ettus/ettus_swig.py",
line 3272, in set_rx_antenna
return _ettus_swig.rfnoc_radio_sptr_set_rx_antenna(self, ant, chan)
RuntimeError: LookupError: KeyError: key "TX/RX" not found in
dict(NSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEE,
NSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEE)
Can anyone advise?
On Wed, Oct 10, 2018 at 9:06 PM Corey Hahn via USRP-users <
usrp-users@lists.ettus.com> wrote:
> Moving to uhd-fpga and checking out 'ebf5eed' did the trick. Recompile
> the fpga. In working order again.
> Thank you for walking be through the rfnoc structure with respect to the
> git repos.
> I still think something is wrong with pybombs to cause this error in the
> first place.
>
> Corey
>
> On Wed, Oct 10, 2018 at 9:48 AM Rob Kossler <rkoss...@nd.edu> wrote:
>
>> Perhaps the issue is related to pybombs. You mentioned that you
>> completely reinstalled via pybombs. I seem to recall that the pybombs
>> installation puts the fpga source into a folder called 'uhd-fpga' or
>> something like that. But, a manual UHD install puts the fpga source into a
>> folder called 'fpga-src'. Whatever folder you are using for the fpga
>> source needs to be on commit 'ebf5eed'. Maybe pybombs is simply checking
>> out 'master' for the fpga source code rather than following the appropriate
>> fpga commit that is associated with the uhd 'master'.
>>
>>
>> On Wed, Oct 10, 2018 at 10:16 AM Rob Kossler <rkoss...@nd.edu> wrote:
>>
>>> I'm confused. If the 'fpga-src' folder is on commit 'ebf5eed', the
>>> compat number should be 35.1. It wasn't changed to 36.0 until the
>>> following commit. I'm pretty sure that I have built an fpga image from
>>> this source and that it ran fine with the uhd 'master' branch.
>>>
>>> Rob
>>>
>>> On Wed, Oct 10, 2018 at 8:41 AM Sylvain Munaut <246...@gmail.com> wrote:
>>>
>>>> > No dice, same ver 35 and 36 mismatch error.
>>>>
>>>> When looking at the commit you mention (
>>>> 6af6ac32c30d2dc0efa6eab61e4a3920649e3e62 ), the expected compat number
>>>> is 35 :
>>>>
>>>>
>>>> https://github.com/EttusResearch/uhd/blob/master/host/lib/usrp/x300/x300_fw_common.h#L26
>>>>
>>>> Looking at the latest FPGA sources on the maint branch, I see also
>>>> compat number 35 :
>>>>
>>>>
>>>> https://github.com/EttusResearch/fpga/blob/maint/usrp3/top/x300/bus_int.v#L104
>>>>
>>>> So you should use the 'maint' branch of the FPGA sources. The latest
>>>> FPGA master seems to have some new down conversion logic for which the
>>>> UHD host code isn't published yet.
>>>>
>>>>
>>>> Cheers,
>>>>
>>>> Sylvain
>>>>
>>> _______________________________________________
> USRP-users mailing list
> USRP-users@lists.ettus.com
> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
>
_______________________________________________
USRP-users mailing list
USRP-users@lists.ettus.com
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
