[En-Nut-Discussion] TCP stops working after some time

[Harald Kipp]

>
>My personal asumption (not checked by reading source) is that our 100ms
>was rounded down to the granularity (~63ms IIRC) and we effectifly
>had a timeout between 0ms and timer granularity because it just waits
>for the next timer interrupt to happen.

That makes a lot of sense to me too.

The granularity is 62.5 ms unless you explicitly set NUT_CPU_FREQ.
If your observations proof to be correct, any timeout less than
125 ms may expire immediately.

Harald