Hang on. He didn't plant the patented seeds; he planted his own seeds, which were then pollinated by his neighbour's Roundup-Ready plants. I can't see anything wrong with that.
So now he has a crop that consists of hybrids. These hybrids are all different; each plant is a different mixture of <farmer's variety> with Roundup-Ready. Some of those plants will have resistance to glyphosate. Then he sprays with Roundup; the resistant plants survive, the rest die. Now he has a harvest of seeds that are resistant to glyphosate, but in all other respects are a mixture. These seeds are not the ones sold by Bayer/Monsanto.
So is it the strain that is patented, or the gene?
If it's the strain that gets the patent, then the farmer-next-door isn't growing your patented strain, so he's in the clear. So it must be the gene, right?
If it's the gene, then it seems unreasonable to yell "patent violation" if you're spreading patented pollen over the entire midwest. You must either sell seeds that don't make (viable) pollen, or you have to accept that the farmer had his crop involuntarily infected with your IP because of your negligence.
I thought the GMO manufacturers bred sterility into their strains for just that reason.
Meanwhile, the farmer-next-door now has a grain-store full of resistant hybrid seeds, that won't (on the whole) fare as well as Roundup-Ready, because they're all different strains. That is, having the resistant gene isn't the whole story; you need a strain that is resistant AND grows well AND is consistent. That means an F1 hybrid, and you don't get that by just crossing strains; you have to get into selecting and cloning the favourable strain, growing that strain into a crop, and letting it cross with its (identical) siblings to make true-bred seeds.
Have I missed something?