AX.100 decoder crashing

Hi,

is there any progress solving this already known issue?

For some time I was able to work around this problem by using gr-satellites, but recently it always crashes with “invalid pointer”. So it is not possible to stable decode any of the ax.100-satellites.

I am using the latest librespace/satnogs-client:master-unstable image

Stefan / OE6ISP

[INFO] Make connection: ‘LimeSDR-USB [USB 3.0] 90726074D061D’
[INFO] Reference clock 30.72 MHz
[INFO] Device name: LimeSDR-USB
[INFO] Reference: 30.72 MHz
[INFO] LMS7002M register cache: Disabled
[INFO] RX LPF configured
[INFO] Rx calibration finished
CCSDS RS Decoder1 :debug: dropping RS block due to error symbols
munmap_chunk(): invalid pointer
netrigctl_close: done
INFO satnogsclient.radio.grsat Stopped gr_satellites
INFO satnogsclient.radio.grsat Processing kiss file
INFO satnogsclient.radio.grsat Total frames: 0
INFO satnogsclient.observer.observer Observation 14694629 finished.

INFO satnogsclient.scheduler.tasks Upload of waterfall image for observati

2 Likes

Not really yet. The only thing I checked is if there was any obvious change to the decoder recently and I don’t see any.

In my opinion the problem is not the decoder itself, it must be something in the environment, like wrong pointer or buffer-alignment that leads to the crash.

Some debug-output at the input and output of the RS decoder would be helpful, I think.

Maybe settings the log level to “DEBUG” it will provide extra information.

Jan | PE0SAT

1 Like

done …

1 Like

there were major changes last october

1 Like

I went back to V2.0 and there it obviousely works perfectly.

So I would assume, it was the update last October, that broke something.

Stefan/OE6ISP

4 Likes

Thanks for checking. I don’t see how changes from that commit affect performance at first glance so I have to check it more thoroughly.

It is not about the performance, but more the fact, that ALL decodes fail with “symol errors” and the observations finally crashing with “invalid pointer” after this commit in 2.1.

CCSDS RS Decoder1 :debug: dropping RS block due to error symbols
munmap_chunk(): invalid pointer

same case with @stefanp , with different satellite:

1 Like

i check on network, and seen all some 2.1.2 station affected by this bug.

indication: the waterfall interval shorter than obs timeframe

1 Like

As far as I could find, the problem is specific for all satellites with AX.100 MODE 5 - as the error-message says, there is a problem with the RS-decoder.

The short waterfall comes from the crash, that terminates the obs early.

2 Likes

Here also a AX.100 obs that fails

Error:

2026-08-10T10:28:28.593619555Z CCSDS RS Decoder1 :debug: decoded RS block with 0 error symbols
2026-08-10T10:28:28.593725927Z munmap_chunk(): invalid pointer

Full log entry:

2026-08-10T10:26:53.002290730Z INFO     satnogsclient.observer.observer Begin observation 14747434 ...
2026-08-10T10:26:53.002500869Z INFO     satnogsclient.radio.grsat       Starting gr_satellites for observation 14747434 at 57600 sps
2026-08-10T10:26:53.845521351Z INFO     satnogsclient.scheduler.tasks   Received job for observation 14747434, starting at 2026-08-10T10:26:53+00:00
2026-08-10T10:26:53.845907934Z WARNING  apscheduler.scheduler           Execution of job "spawn_observer (trigger: date[2026-08-10 10:26:53 UTC], next run at: 2026-08-10 10:26:53 UTC)" skipped: maximum number of running instances reached (1)
2026-08-10T10:26:54.114818720Z INFO     satnogsclient.observer.observer Start gnuradio thread.
2026-08-10T10:26:54.254158488Z ZMQ IQ SUB Connecting to 127.0.0.1:16888
2026-08-10T10:26:54.295264451Z ZMQ FRAME SUB Connecting to 127.0.0.1:16887
2026-08-10T10:26:54.307344517Z ZMQ audio SUB Connecting to 127.0.0.1:16889
2026-08-10T10:26:54.436439756Z rig_check_rig_caps: p1=0x7f92bceb1c40, p2=0x7f92bceb81c0, rig_model=0x7f92bceb1c40, macro_name=0x7f92bceb81c0
2026-08-10T10:26:54.436461325Z initrigs4_dummy: _init called
2026-08-10T10:26:54.436711103Z rig_init: rig_model=Hamlib NET rigctl
2026-08-10T10:26:54.436850794Z rig_init: rx_range_list1 is empty, using rx_range_list2
2026-08-10T10:26:54.437001570Z rig_init: rig does not have tx_range!!
2026-08-10T10:26:54.437147049Z netrigctl_init version 20230106.0
2026-08-10T10:26:54.544778507Z [INFO] Using format CF32.
2026-08-10T10:27:53.765674906Z INFO     satnogsclient.scheduler.tasks   Upload of waterfall image for observation 14745946 successful.
2026-08-10T10:27:54.503386732Z INFO     satnogsclient.scheduler.tasks   Upload of audio recording for observation 14745946 successful.
2026-08-10T10:28:28.593619555Z CCSDS RS Decoder1 :debug: decoded RS block with 0 error symbols
2026-08-10T10:28:28.593725927Z munmap_chunk(): invalid pointer
2026-08-10T10:28:30.132701448Z netrigctl_close: done
2026-08-10T10:28:33.040762758Z INFO     satnogsclient.radio.grsat       Stopped gr_satellites
2026-08-10T10:28:33.040873106Z INFO     satnogsclient.radio.grsat       Processing kiss file
2026-08-10T10:28:33.040967050Z INFO     satnogsclient.radio.grsat       Total frames: 0
2026-08-10T10:28:33.041122055Z INFO     satnogsclient.observer.observer Observation 14747434 finished.
2026-08-10T10:28:33.041142966Z INFO     satnogsclient.observer.observer Executing post-observation script.
2026-08-10T10:28:37.512220169Z INFO     satnogsclient.observer.observer Artifacts upload for observation 14747434 successful.
2026-08-10T10:30:54.208071294Z INFO     satnogsclient.scheduler.tasks   Upload of waterfall image for observation 14747434 successful.
2026-08-10T10:30:54.651269473Z INFO     satnogsclient.scheduler.tasks   Upload of audio recording for observation 14747434 successful.

Looking at the recorded IQ file size, I get the impression the whole flowgraph is crashing

Jan | PE0SAT

2 Likes

… the log looks exactly like mine.

My station reported all records with wrong RS block and usually crashed after the 5th, sometimes earlier.

1 Like