evl ps cpu time reporting issue

Hy there,
I found a few issues with the output of ps -st.
The most crucial one is that sometimes it reports excessive CPU usage, e.g. (running watch -n1 evl  ps -st):


  CPU   PID   ISW     CTXSW     SYS       RWA       STAT     TIMEOUT      %CPU      CPUTIME       NAME
    1   947   105     119       203       1          X          -           0.0      0:011.988    test-pd:947:1804289383
    1   952   2       2         3         0          X          -           0.0      0:000.087    systemTask
    1   954   2       2         3         0          X          -           0.0      0:000.056    WSServer_40100
    1   956   2       2         3         0          X          -           0.0      0:000.070    WSClient_signalMonitor40100_bin
    1   1003  2       2         3         0          X          -           0.0      0:000.054    WSServer_5555
    1   1004  2       2         3         0          X          -           0.0      0:000.054    WSClient_gui_data5555_bin
    1   1005  2       2         3         0          X          -           0.0      0:000.044    WSClient_gui_control5555_bin
    1   1006  15      15        47        0          X          -           0.0      0:000.594    WSServer_5432
    1   1007  504     514       505       0          X          -           0.0      0:021.035    WSClient_scope_data5432_bin
    1   1008  2       2         3         0          X          -           0.0      0:000.049    WSClient_scope_control5432_bin
    1   1009  48842   73393     123151    24420      W          -           0.0      0:380.085    scope-trigger-task
    0   1011  6640    31157     37608     23656      W          -         1193652.1      0:642.101    thread:947:0xffffb3c0fe30:2214587712
    1   1012  774278  1546216   776625    0          D       001.393        0.5      7:690.643    libpd-fdTask
    1   1013  0       2         4         0          W          -           0.0      0:000.081    touchSensorRead
    3   1019  2       28822348  28885721  27                 011           48.3   1092:295.224    bela-audio

note that thread:947:0xffffb3c0fe30:2214587712 has 1193652.1% CPU. In another case it reported 1405727.9, in another something like 16000; some other times it shows between 60% and 150% ... In the instances I've seen, it was only this thread that was affected, no other one. The thread should never go anywhere 10%: most of the time it sleeps on an evl_wait_event(); when it's woken up, it reads 4k of data from disk (thus switching to secondary mode), puts it in a buffer, signals another evl condition variable via evl_signal_event() (thus switching to primary mode) and goes back to sleep in evl_wait_event().

Now, I cannot exclude that there may be something wrong in my code, which may justify the occasional 60% reading  (I didn't write it myself, only ported it to evl, and I didn't inspect it closely), but there's clearly something pathological in evl's CPU reporting system with anything above 100%, I guess?

As usual, I am running this on ARM64 (PocketBeagle 2)
evl.0.55 -- #98a8b88 (2025-10-08 22:41:39 +0000) [requires ABI 42]
Linux bela 6.12.49-arm64-ti-evl #4 SMP EVL Sun Nov  2 19:24:32 UTC 2025 aarch64 GNU/Linux

Best,
Giulio