The chromosome map and the McKernen paper
discussing a stable intermediate line, are convincing lines of evidence (to me at least) to support that that the hypothesis that THC/CBD determination was a single, mutually exclusive locus is wrong, and there are two distinct, nonmutually exclusive genes. But they, and I, could be wrong. Regardless the crosses act as if it’s a single locus, so for all but the most rare circumstances (and definitely for the purpose of the original question asked in this thread), you can assume it will act as a single locus.
In the first generation (F1) of a CBD dom and THC dom cross ALL will be 1:1 (unless, say there is a super rare null, recessive B-allele but this can be assumed not to be present unless you know it is in your genepool (ie CBG plants when homozygous)).
Codominance is fully in line with simple Mendelian genetics and Punnet squares. THCAS and CBDAS are codominant when expressed together. They compete for, and share the CBG precursor, with some variation for enzyme affinity or possibly expression levels.
Linked genes are are a known exception to the Law of Independant Assortment. Why wouldn’t you account for that since we know these are linked?