I've found this to be an interesting question with highly respected monastics falling on both sides of the answer.
Actually, a point that is not well understood is that, among those who follow the Pali canon, there is no controversy over whether some jhāna is necessary; the answer is in all cases yes. The controversy is really over which jhāna is necessary.
There are three types of jhāna in total:
- samatha-jhāna - jhāna based on a conceptual object that, due to its conceptual nature, cannot lead to the realization of the nature of reality.
For example, in the Visuddhimagga:
The colour should not be reviewed. The characteristic should not be given attention. But rather, while not ignoring the colour, attention should be given by setting the mind on the [name] concept as the most outstanding mental datum, relegating the colour to the position of a property of its physical support. That [conceptual state] can be called by anyone he likes among the names for earth (pathavī) such as “earth” (pathavī), “the Great One” (mahī), “the Friendly One” (medinī), “ground” (bhūmi), “the Provider of Wealth” (vasudhā), “the Bearer of Wealth” (vasudharā), etc., whichever suits his manner of perception. Still “earth” is also a name that is obvious, so it can be developed with the obvious one by saying “earth, earth.” It should be adverted to now with eyes open, now with eyes shut. And he should go on developing it in this way a hundred times, a thousand times, and even more than that, until the learning sign arises.
2. vipassana-jhāna - jhāna based on ultimate reality that allows one to see the three characteristics and thus attain nibbana. As the Mahasi Sayadaw states:
The tranquillity that occurs while one contemplates various phenomena is called momentary concentration, the concentration that lasts momentarily during contemplation. No insight is possible without this momentary concentration. The meditator who has no basic [samatha] jhānic experience and relies on insight meditation alone develops insight through momentary concentration and attains the noble path. This concentration for insight is not confined to a single object. The meditator practising it notes all the mental and physical phenomena that arise. However, at the moment of noting, his mind is fixed on the object and free from distraction. This is obvious to the meditator who practises effectively.
3. lokuttara-jhāna - jhāna that takes a supermundane object (nibbana), and is equivalent to attaining the noble path and fruition. According to Bhante Gunaratana:
But the four jhanas again reappear in a later stage in the development of the path, in direct association with liberating wisdom, and they are then designated the supramundane (lokuttara) jhanas. These supramundane jhanas are the levels of concentration pertaining to the four degrees of enlightenment experience called the supramundane paths (magga) and the stages of liberation resulting form them, the four fruits (phala).
It is this third type of jhāna that everyone agrees is necessary. The problem is with deciding which type of jhāna the Buddha is talking about in specific contexts. Since these three types of jhāna are not enumerated in the sutta pitaka, the debate centres around whether lokiya (mundane) samatha jhāna are necessary, since that is the only type of jhāna recognized by certain Buddhists.
This too can be seen as a misunderstanding, since samatha jhāna according to those who hold that it is not necessary is not exactly the same as jhāna according to those who hold that lokiya jhāna is.
According to those who say lokiya jhāna is necessary, they seem to believe that lokiya jhāna is a practice which allows for the realization of insight, in which case it cannot be samatha jhāna as understood by those who believe that samatha jhāna cannot lead directly to insight, unless the former group actually believes that jhāna based on a conceptual object can allow for the realization of insight into reality, which I'm pretty sure is not the case.
So, the debate is really not all it is cracked up to be, unless there are people who believe meditation based on a non-real conceptual object is sufficient for attainment of understanding of reality, in which case they are in contradiction with the sutta pitaka, not to mention the abhidhamma and commentaries.
I suppose there could still be room for debate over whether samatha jhāna based on a concept must be cultivated prior to Vipassana; the yuganaddha sutta probably shows this to be false:
Then there is the case where a monk has developed tranquillity preceded by insight. As he develops tranquillity preceded by insight, the path is born. He follows that path, develops it, pursues it. As he follows the path, developing it & pursuing it — his fetters are abandoned, his obsessions destroyed.
Of course for anyone actually adhering to the Theravada (I.e. the commentaries, patisambidhamagga, and abhidhamma), there isn't much room for controversy. Most of the debate is over whether to follow the existing commentaries or create one's own.
Well, I think there is a reason why the path is eightfold and not sevenfold. The Pāḷi Canon is full of encouragements, made by the Buddha, to practice jhānas. It would be stretching the texts a lot to say it isn't necessary. I wouldn't argue over it though.
‘The sage, the withdrawn chief bull,
the Buddha who awakened to jhāna,
the One of Broad Wisdom has found
the opening amid confinement.’
-AN 9.42, Confinement (Sambādha-suttaṃ) & SN 2.7, Pañcālacaṇḍa (Pañcālacaṇḍa-suttaṃ)
'There's no jhāna for one with no discernment,
no discernment for one with no jhāna.
But one with both jhāna & discernment:
he's on the verge of Unbinding.'