Could a bacteriophage package an entire host genome by generalized transduction?
It's not impossible. Generalized transduction isn't part of the usual phage life cycle but it's one of the ways phage can mediate horizontal gene transfer. There's a limit on how much genetic material transduction can transfer as the phage capsid is only so large. That being said, some phages carry rather large genomes. The bacillus phage G genome is one of those hefty ones at nearly 500 kb and almost 700 open reading frames. It was found in Bacillus megaterium. Depending upon the strain, a B. megaterium genome is between 5.1 and 5.7 Mb, or at least ten times the size of the phage G genome.
That being said, the Mycoplasma genitalium genome is only around 580 kb. I don't know if any phages are known to infect it but one as large as bacillus phage G could hypothetically package a nearly complete M. genitalium genome. If we consider endosymbionts like Carsonella ruddii or Nasuia deltocephalinicola in this scenario then their < 200 kb genomes would even fit into smaller viral capsids. How would the phages enter a symbiont-containing host in the first place?
I suspect that this phenomenon is much more likely to occur in a lab than in the wild but phages have a way of subverting our expectations.