The topic is discussed quite thoroughly by Kātyāyana and Patañjali in the Mahābhāṣya ad P. 2.1.1 (and of course the various sub(sub) commentaries thereon).  The discussion is technical of course, and the pros and cons of each mode of analysis are hashed out at length.  In the translation of the āhnika by Joshi and Roodbergen (Patañjali's Vyākaraṇa-Mahābhāṣya Samarthāhnika 1968) the discussion begins on p. 168 with Patañjali's prelude to Vārtt. 25. The answer to your question, though, is stated plainly in the translators' introduction (p. xvii): "The conclusion that they arrived at is that dvandvas or a bahuvrīhi can be formed of more than two constituent words. Elsewhere a compound is generally formed of two words only."

This analysis is also reflected in commentators' glosses. See Scholastic Sanskrit by Tubb and Boose p. 126 for examples.

See also the Kāśikāvṛtti ad P. 2.2.29 cārthe dvandvaḥ where the gloss of dhava-khadira-palāśāḥ is given as dhavaś ca khadiraś ca palāśaś ca. Note that the word enekam is continued from 2.2.24 which treats bahuvrīhi compounds and is used to indicate that such compounds can be formed from multiple words not just two as is usually the case by sup supā etc.

All the Best,

Victor