The gemara discusses whether workers continue working while they say shema, and concludes that for the first perek they stop their work to say shema, but the second chapter they say while working. The gemara is assuming that the first perek requires kavana, but the second does not. The issue of kavana was already dealt with at the beginning of the perek regarding the part that one has to stop walking for. Rava holds only the first pasuk requires intent, whereas R. Yochanan says the entire first parsha. The Ri"f explains that even Rava who only requires kavana for the first pasuk and would allow one to continue walking after the first pasuk, would say that one cannot do work for the entire first parsha דלא לשוי לה עראי - not to degrade the first parsha of shema. However, Tosafos says that the gemara's answer is to allow for a distinction even in the opinion of R. Yochana, but since we pasken like Rava it would be sufficient to stop working for just the first pasuk.
Rabbeinu Yona points out that according to the Ri"f the d'oraysa part of shema is only the first pasuk, yet we force workers to stop their work for the first parsha דלא לישוי עראי, but after that we are not concerned. Why are we not concerned after the first parsha? From here the Rabbeinu Yona derives that only the first parsha of shema is d'oraysa, the parsha of והיה אם שמוע is only d'rabonon. Rabbeinu Yona elaborates to prove this point. One of his sources is the gemara 13b that Rebbi Yehuda HaNasi would only say shema yisroel within the time of Krias Shema. Rabbeinu Yona understands this to mean that he would read the entire first perek, not just the first pasuk. If the second perek were d'oraysa Rebbi would have had to say that as well, the fact that he didn't proves that after the first perek it is only d'rabonon.
In the gemara on 13b about Rav Nachman who would dose off during shema and only required his servant to keep him awake for the FIRST PASUK, not the first parsha. If the first parsha is d'oraysa why didn't he require his servant to keep him awake for the entire parsha? Rabbeinu Yona (beginning of perek) explains that he actually read the entire shema, just that he was dosing off so he didn't have kavana. He insisted that his servant keep him alert to have kavana for the first pasuk, like Rava paskens, but not beyond that.
This distintion between the first parsha being d'oraysa and second being d'rabonon, would help explain the opinion of Rashi on 2a who says that the mitzvah of krias shema is fulfilled by the saying of the first perek before one goes to sleep. Tosafos disagrees and says that if one were to be yotzei with krias shema al ha'mita, they should read all three parshiyos (the third is definitely not d'orasya for shema, just for yetzias mitzrayim). The Sha'agas Aryeh (3) also justifies the position of Rashi based on the Ri"f that only the first parsha is d'oraysa, but for the d'rabonon parsha we rely on the saying of shema early. The proof of Rabbeinu Yona has that the second parsha isn't d'oraysa is based on his learning of our gemara that the distinction between the first perek and second is true even according to Rava who requires kavana only for the first PASUK, so the distinction between the first PARSHA and second must be whether it is d'oraysa. However, according to Tosafos that the workers only need to pause for the first PASUK, that is because of kavana, but the d'oraysa requirement to read could be both the first and second PARSHA, which is li'shitasam on 2a.
Rabbeinu Yona asks that since we require people learning to stop for all mitzvos, even Rabbinic mitzvos, why wouldn't we require Rebbi to stop for והיה אם שמוע? Rabbeinu Yona cites a Yerushalmi that we stop a reading mitzvah i.e. limud ha'torah, for an action mitzvah, but not for another reading mitzvah i.e. krias shema.