Since I'm working with the database errors #911 and #912 for incorrect patronymics, Eric has been on my to-do-list since October - he should have Ersson as LNAB.
Antonia is right that matronymics are very uncommon. I have never come across any in my own kinship network; there are always other solutions for kids born out of wedlock. But you can find them if you look for them. I explored the matter a bit and have come to the conclusion that it was done mostly in the strictly religious parts of Sweden = Bohuslän (schartauanism) and Norrland (laestadianism). So I view it as a form of public shaming.
The root cause of the problem of LNAB's of children in the Swedish past is that they did not really HAVE a last name at birth. I think it was Ingela Martenius who wrote that children were just the property of their parents/their father. In birth records it is very rare that children are mentioned by Firstname Lastname. We just assume that they have the patronymic that they will be known by later in life (or family name if there is one). Children living with their parents are almost never in the household records with a last name (excepting stepchildren).
For children out of wedlock, when the father was known, the child was very often named after him (patronymic or family name). Indeed, now that the local gossip of the time is forgotten, this name may be the first clue to the identity of a Father Unknown.
When the father was genuinely unknown (or kept a secret) the solution would be as in Eric's case: to base a patronymic on the maternal grandfather's name (or take his family name, without a patronymic).