Khmer script is used for writing different languages : khmer (in its different flavours) but also some minority languages, as well as pali, sanskrit, thai and lao.
This seems covered by the current unicode specifications, except for it might be not precise enough. Especially about the order of subscripts, diacritics and dependant vowels.
In their response to Cambodian official objection to Khmer block, Maurice Bauhahn and Michael Everson suggest using variation-selector-1 for the alternate viriam shape but it is missing from the official list. This would allow typing the article about viriam (វិរាម) in the Chuon Nath dictionary.
A variant sequence for ឳ would also allow encoding Guesdon's Dictionnaire Cambodgien-Français.
A mirrored version of ៚ could also be useful as it is used a lot in poetry at the beginning of a verse (this could also be achieved by font features).
According to the EFEO's paleographic tables, two 6th century's consonants as well as one 10th century's diacritic are missing. There were also special signs for many numerals which I guess should be added too... The syllable structure would need to be adapted too as it seems to have been much more flexible by then...
As far as I know, the current unicode standard fully supports pali and sanskrit in khmer script.
Khmer script (aka khom script) has been extensively used in thaï religious texts, and to a certain extent in lao as well (see AKSOON KHOOM: Khmer Heritage in Thai and Lao Manuscript Cultures by Jana Igunma).
In order to fully support thaï in khmer script, one would need the khmer characters corresponding to the thaï ◌้ (U+0E49 THAI CHARACTER MAI THO) and ◌๊ (U+0E4A THAI CHARACTER MAI TRI). For the characters and , they would need to be added as well, or be considered as ligatures but in this case the rule should be changed to allow vowels after U+17CD (see Michel Antelme's Inventaire provisoire des caractères et divers signes des écritures khmères [...]).
The addition of 2 characters for typesetting Krung language was "postponed for future study" back in 2002 (see L2/02-099 and L2/02-291). These were removed from the official orthography in 2003 (see official documents here and here).
Knowing wether the size of the existing corpus using these signs makes it worth adding them to the standard is way beyond my expertise but I list it here for the sake of completeness.
According to M. Danh Hong, there are also characters missing in order to typeset Surin language.