One GFCI downstream of another is unnecessary and makes troubleshooting difficult.
The current leaving the GFCI load hot terminal has to equal within 5 mA the current returning on the GFCI load neutral terminal.
In principle, GFCIs only know the difference current sensed by the internal toroidal transformer [see fig. 2 below]
and are insensitive to voltages.
My wife's hair dryer has a GFCI built into the plug but it's always the wall GFCI that trips, after a few minutes.
No hair dryer, no nuisance trips. They happen infrequently which makes troubleshooting almost impossible.
One of these days I'll salvage the toroid out of the failed GFCI in my basement junkbox and use it to actually measure the differential current. This is a lot cheaper than buying Fluke's instrument that does this.
With Romex you can typically expect a half mA of current to ground due to inter-conductor capacitance.