Flight Pro 820 - DSPAL tester failed

[plusk01]

I would like to use PWM on the new Flight Pro board.

After following the instructions from Intrinsyc, I can use dspal/test to get the version:

cpu

root@apq8096/home/root $ ./version_test 
version: 2.0.0
build date: BUILD_DATE_STRING=Oct 22 2017
build time: BUILD_TIME_STRING=22:12:45

dsp

[08500/02]  41:23.050  Verification skipped, no function specified!!  0495  map_object.c
[08500/02]  41:23.050  uimg not supported  0529  map_object.c
[08500/02]  41:23.071  Verification skipped, no function specified!!  0495  map_object.c
[08500/02]  41:23.071  uimg not supported  0529  map_object.c
[08500/03]  41:23.083  __cxa_finalize weak ref not found, return _rtld_sym_zero@_rtld_objmain  0295  symbol.c
[08500/03]  41:23.083  __cxa_finalize weak ref not found, return _rtld_sym_zero@_rtld_objmain  0295  symbol.c
[08500/03]  41:23.083  __cxa_finalize weak ref not found, return _rtld_sym_zero@_rtld_objmain  0295  symbol.c
[08500/03]  41:23.083  __cxa_finalize weak ref not found, return _rtld_sym_zero@_rtld_objmain  0295  symbol.c
[08500/02]  41:23.086  version: 2.0.0  0040  version_test.c
[08500/02]  41:23.086  build date: BUILD_DATE_STRING=Oct 22 2017  0041  version_test.c
[08500/02]  41:23.086  build time: BUILD_TIME_STRING=22:12:45  0042  version_test.c

but the dspal_tester fails with (except HAP power API test):

dsp

// ...
[08500/03]  41:46.193  undefined PLT  symbol #489 __cxa_atexit in libdspal_tester.so  0318  symbol.c
[08500/03]  41:46.193  plt object relocation failure  0274  mdreloc.c
[08500/03]  41:46.195  dlopen failed, libdspal_tester_skel.so  0725  rtld.c
[08500/02]  41:46.215  Verification skipped, no function specified!!  0495  map_object.c
[08500/02]  41:46.215  uimg not supported  0529  map_object.c
[08500/02]  41:46.242  Verification skipped, no function specified!!  0495  map_object.c
[08500/02]  41:46.242  uimg not supported  0529  map_object.c
[08500/03]  41:46.258  __cxa_finalize weak ref not found, return _rtld_sym_zero@_rtld_objmain  0295  symbol.c
[08500/03]  41:46.260  __cxa_finalize weak ref not found, return _rtld_sym_zero@_rtld_objmain  0295  symbol.c
[08500/03]  41:46.260  __cxa_finalize weak ref not found, return _rtld_sym_zero@_rtld_objmain  0295  symbol.c
[08500/03]  41:46.261  __cxa_finalize weak ref not found, return _rtld_sym_zero@_rtld_objmain  0295  symbol.c
[08500/03]  41:46.261  undefined PLT  symbol #489 __cxa_atexit in libdspal_tester.so  0318  symbol.c
[08500/03]  41:46.261  plt object relocation failure  0274  mdreloc.c
[08500/03]  41:46.261  dlopen failed, libdspal_tester_skel.so  0725  rtld.c
[08500/02]  41:46.283  Verification skipped, no function specified!!  0495  map_object.c
[08500/02]  41:46.283  uimg not supported  0529  map_object.c
[08500/02]  41:46.311  Verification skipped, no function specified!!  0495  map_object.c
[08500/02]  41:46.311  uimg not supported  0529  map_object.c
[08500/03]  41:46.330  __cxa_finalize weak ref not found, return _rtld_sym_zero@_rtld_objmain  0295  symbol.c
[08500/03]  41:46.330  __cxa_finalize weak ref not found, return _rtld_sym_zero@_rtld_objmain  0295  symbol.c
[08500/03]  41:46.331  __cxa_finalize weak ref not found, return _rtld_sym_zero@_rtld_objmain  0295  symbol.c
[08500/03]  41:46.331  __cxa_finalize weak ref not found, return _rtld_sym_zero@_rtld_objmain  0295  symbol.c
[08500/03]  41:46.332  undefined PLT  symbol #489 __cxa_atexit in libdspal_tester.so  0318  symbol.c
[08500/03]  41:46.332  plt object relocation failure  0274  mdreloc.c
[08500/03]  41:46.332  dlopen failed, libdspal_tester_skel.so  0725  rtld.c
[08500/02]  41:46.355  Verification skipped, no function specified!!  0495  map_object.c
[08500/02]  41:46.355  uimg not supported  0529  map_object.c
[08500/02]  41:46.383  Verification skipped, no function specified!!  0495  map_object.c
[08500/02]  41:46.383  uimg not supported  0529  map_object.c
[08500/03]  41:46.391  __cxa_finalize weak ref not found, return _rtld_sym_zero@_rtld_objmain  0295  symbol.c
[08500/03]  41:46.391  __cxa_finalize weak ref not found, return _rtld_sym_zero@_rtld_objmain  0295  symbol.c
[08500/03]  41:46.393  __cxa_finalize weak ref not found, return _rtld_sym_zero@_rtld_objmain  0295  symbol.c
[08500/03]  41:46.393  __cxa_finalize weak ref not found, return _rtld_sym_zero@_rtld_objmain  0295  symbol.c
[08500/03]  41:46.393  undefined PLT  symbol #489 __cxa_atexit in libdspal_tester.so  0318  symbol.c
[08500/03]  41:46.393  plt object relocation failure  0274  mdreloc.c
[08500/03]  41:46.395  dlopen failed, libdspal_tester_skel.so  0725  rtld.c

Is this related to commit 291a68c? What is SLPI vs DSP?

See also: QDN forum.

[edwinpha]

@plusk01 could you use PWM with the new board?

[plusk01]

Hi @edwinpha, unfortunately the sfpro does not have GPIO configurable for PWM. Instead, we have been using a PCA9685 I2C to PWM board for sfpro->esc communication.

See here for more information.

[ilNonn0]

Hi @plusk01 thanks for sharing the link above.
Just a question: do you know which is the command sent from the sfpro to the Qualcomm ESC board by UART channel?
My doubt is the following: I saw the Qualcomm Snapdragon Navigator ESC
Protocol Specification document here where it is suggest to use RPM command for flight (pag 6).
Now.. If I replace the Qualcomm ESC board with a custom one and I use the the PCA9685 I2C to PWM board to bridge the two components (sfpro and custom ESC board), is it possible to use the original Qualcomm Navigation software? I don't think so since it is based on a closed loop RPM control.
Do you confirm?
How did u manage that?

[plusk01]

@ilNonn0 we do not use snav package; instead, we have our own custom autopilot which outputs PWM commands. We do not use the ESC board since we have vehicles with more than four motors, so I am not sure about your question, but your assessment seems logical.

[ilNonn0]

@ilNonn0 we do not use snav package; instead, we have our own custom autopilot which outputs PWM commands. We do not use the ESC board since we have vehicles with more than four motors, so I am not sure about your question, but your assessment seems logical.

@plusk01 that's sound good
Maybe in the next days, we will switch to it as well since Qualcomm documentation seems too vague and we need to customize some aspects.
Just in case we will adopt your software, is it fine for the sfpro as well right?

[plusk01]

@ilNonn0 that's correct. We have a variety of vehicles using either the sfpro or old sf.