The SystemVerilog spec describes a scheduler model in Figure 4-1 (p. 67 of IEEE 1800-2023). It describes a number of regions, including an Active region set and a Re-Active region set.
Orthogonally, it also describes a set of dedicated PLI regions, some of which are in the Active set, and some in the Re-Active set.
The spec then goes on to describe PLI callback control points in Table 4-1 of Section 4.10 (p. 72). This table describes 7 PLI backs (which are actually VPI callbacks in Chapter 38), and their mapping to event regions.
The mapped regions, however, are all in the Active region set (and the Pre-Postponed and Postponed regions). There are no callbacks that map to any Re-Active region.
So how does a VPI application schedule things in any Reactive region? Indeed, how is anything scheduled in the Reactive PLI region set? I cannot find any reference elsewhere in the spec.
The LRM is missing this due to lack of volunteers to address this section of VPI specification. They remain as open issues. here is a link to one of them