If you think of language as a whole, then 'knowledge about language' is made up of anything and everything that describes language or can explain why and how we use it in the ways that we do.
Over the last few years, 'knowledge about language' in the hands of the government, the DfE and Michael Gove has been reduced to 'grammar' and 'grammar' has been reduced to one model, one form of what 'grammar' might be - a so-called 'structure and function' model.
This single model of 'grammar' (treated as if it's the only model) and enforced through the Grammar, Punctuation and Spelling test, then holds sway over primary education, and primary aged children.
First, to be clear, there are other models of grammar, which, say, treat that word 'function', not as how words 'function' inside sentences (e.g. this noun is the subject of the sentence) but as social functions (e.g. why have so many of us started saying 'So...' at the beginning of our utterances).
For some reason, this form of grammar was not the one implemented and enforced.
There is, though, an even more important criticism to make. 'Knowledge about language' is a massive subject and can't be reduced to 'grammar' of any kind. Since the time of Aristotle, linguists have tried to examine language, describe it and explain it. Aristotle was particularly interested in the 'effects' of particular uses of language and did a damned good job of it. We all know, for example, what 'catharsis' is, thanks to him, but he did more than that in his book 'Poetics'.
Over the last 150 years, a huge amount of work has gone into examining how the many different uses of language work and have created disciplines such as narratology, stylistics, pragmatics and intertextuality. Though these are mostly written about in very academic ways, they can be broken down into very accessible (and enjoyable) ways for children and school students to use. To be clear: these are also 'knowledge about language', and because they are tied very closely to 'language in specific uses' and not 'abstract ideals', they are especially useful in helping children speak and write.
Narratology, for example, enables us to examine how stories (or any kind of writing) are 'told': e.g. who narrates? how does the narration change? what kind of narrator is narrating? what devices does the narrator use to 'talk' to us?
Narratology can help us look at how the narration enables us to know how characters think. There are several very different devices that have grown up, all the way from 'she thought' to the 'free indirect discourse' favoured by Jane Austen and many writers of children's books.
Narratology can help us look at 'foregrounding' and 'point of view' - how these shift, favouring one or more characters and why?
Narratology is very useful at helping us with time frames which often change via flashback, flash forward and invocations of continuous time or continuous existence.
Stylistics can take us into how texts 'sound' (prosody) - showing us how repetition of structure and letter sounds make rhythms in texts.
Stylistics can draw attention to sentence length, sentence complexity or simplicity, how paragraphs are constructed across texts, why and how these change as the need to express different things change.
Stylistics can draw attention to 'register' - how informal/formal a text is? How much does it draw on modes of text from which sources - does the writing empty speech modes? Are there deliberate attempts to 'borrow' language from specific sources e.g. from a field different from the one in the text, e.g. from science in a novel?
Stylistics can draw attention to which class of words are repeated e.g. many adjectives, many adverbs - or none?
Pragmatics can draw attention to how dialogue is structured and where the narrator dialogues with the audience/readership. Dialogue can be structured in many different ways in fiction and pragmatics can help us make distinctions.
Intertextuality can help us with the matter of 'borrowing' that I mentioned earlier. In essence, all writing is borrowing in that it borrows the sounds, structures and meanings that have gone before in order to do whatever it does. However, some borrowings are more obvious than others and/or more significant. This can be at the level of a whole genre e.g. Hamlet as 'revenge tragedy' or at the level say of using literary motifs or tropes e.g. 'the pathetic fallacy'. Or again allusion to writing or speech that comes before (as Dickens does in the opening pages of 'A Christmas Carol') and so on.
If the government and the DfE had been really interested in a holistic view of language and 'knowledge about language' it would have talked to applied linguists about all this, and then got hold of people who know about pedagogy and asked them to produce materials which applied this 'knowledge about language' in age-appropriate ways, using imitation, and practice and investigation as much as description and direct instruction, so that this 'knowledge about language' could have been applied directly to helping children write well.
But they didn't.
The main reason why they didn't is because the Bew Report of 2011 imposed the SPaG test instead. This was because Michael Gove told them to.