Yes, this is a well-known rule (see Leumann 1977: 132-133, Tronskii 1960: 101, Weiss 2009/2011: 124 and many others).
Weiss puts it this way,
“An unstressed short vowel between a liquid and u̯ is syncopated” (p. 124):
*bholh1i-u̯os > *foliu̯os > fulvus
*sl̩h₂u̯os (or *sl̩h₂euos) > *salau̯os > salvus
*arau̯om > arvum
cf. Leumann "Postkons. lat. u̯ (v) vor Vokal aus vokal. u: vorhistorisch hinter r l; hinter anderen Konsonanten aus Versnot bei Daktylikern, spaeter allgemein im Vulgaerlatein" (p. 132).
It occurred in the pre-literary period.
Exception:
“If the vowel in the preceding syllable was long, the syncope did not occur until the post-Plautine period:
mīluus > mīlvus
lārua > lārva
*pēleu̯is > pēluis > pēlvis