Could I assume that every subhalo hosts a single galaxy and, therefore, that I can interpret the 6 entries of the "SubhaloMassType" field as the masses of all types of particles in each single galaxy?
And could I also assume that the first entry of this field is the gas mass and that the other ones follow the GADGET naming convention?
This is broadly right, for instance in almost all the figures of the intro papers, "galaxies" does just mean subhalos.
And the "*Type" fields (with 6 entries) are the same numbering as the snapshots (0=gas, 1=dm, 4=stars+wind, 5=bhs).
Two complications you might want to keep in mind: