Pests that directly attack a crop (eg, insects, mites, fungus) might be controllable via an engineered resistance gene (or genes). However, weeds are a different matter. Biological control of weeds in crops hasn't really been very effective, which leaves chemical or mechanical control. Among herbicides, glyphosate has tested out as about the least toxic and persistent. Being able to kill just about ALL competing plants in a field, with hardly any effect on the crop must have seemed like a miracle when it was first released. Now that glyphosate is off-patent, competitive manufacturing has driven down the price, making the RR crop/glyphosate combination an irresistable tool for many farmers.
Only if the market (ultimate consumers or supply chain) suddenly starts to shun these crops will farmers turn away from them because they currently offer significant business advantages to the farmers' operations.
However, as with all control methods (to a degree, even mechanical), Nature Bats Last. Resistance management may delay the inevitable, but years of use strongly selects for resistance. Glyphosate resistant weeds are showing up in increasing numbers already, forcing the use of more complicated control methods (eg multiple chemical/mechanical combinations), and similar pressures will eventually undermine the effectiveness of built-in pest resistance, too (eg Bt crops or virus-resistant papaya).
Final consumer traits, like those found in Golden Rice, Arctic Apple or Innate Potato shouldn't face the same selective pressures and might be expected to have extended useful field lifespans as a result.