A plane stalls from high AoA, not from IAS.
Derived IAS from GS is not really relevant.
The very high initial ROC of 6000 fpm points to a high AoA imho.
Cleaning the config may contribute.
The afore mentioned data shows a steady shallow climb and acceleration consistent with the phase of flight and the previous go around. There is no evidence of anything other than an adequately conducted manoeuvre. If they were in a relatively steady state flight path, which the data shows, and increasing airspeed, which the data shows, and maintaining the track, which the data shows, then they cannot have stalled. The sample rate of the FR24 data would be sufficient to show evidence of a stall, if it were to exist.
It's never been a problem, but then how many times has an aircraft actually held for 2 hours in icing conditions?
Actually they were vectored for an hour and held for an hour. They requested FL80 after a missed approach. The obvious inference is at FL80 they would be in VMC and hence not in icing conditions. Vectoring was at FL80 and holding at FL150. So it would seem they were not in icing conditions for anything other than the approach and missed approach phases. I think the chances of icing being a factor as extremely unlikely.
But the video shows they were almost diving vertically at that point
With a bit of trig, the descent angle was approximately 45 degrees and the avg speed in the descent around about 320kts. That would be consistent with the videos.
So I would say there is no evidence of uncommanded reverser, stall, uncommanded roll, mishandling upto the descent, significant icing nor bad judgement. The current FR24 data indicates only a very sudden pitch down with no other changes in parameters or their trend. What is most likely to be an uncommanded pitch down.