If I have low-side shunt, I would not be able to detect short-circuit, so that goes out of question.
Voltage clump point is simple: I have Zener (15V) connected between gate and source. Gate is connected to a simple mosfet switch (powered from a separate voltage of +49V, so I have 6V to fully open my H-Bridge high side mosfet). If my +43V driving voltage for a bridge would stay permanent, this would be no problem then: source is either at ~+43V and gate at +49V (6V difference), or source is at ground together with gate (~0V difference). But my +43V can go anywhere between +43V and zero. So, when high side mosfet is on (gate at +49V), source can be, for example, at +15V. So, this difference needs to be clamped. Zener does clamp perfectly, except it starts to infuse current into the load; this current is taken from +49V rail of driving mosfet, goes into the load via Zener, and bypasses my current sensing resistor. I can reduce this extra current by increasing resistor between +49V rail and drain of driving mosfet, but a) it will never be small enough and b) this will reduce switching time.
It is also my understanding, that most (all?) high side driver IC use Zener clamps in this exact config, so this would not solve the issue. Or are there any IC that use different approach?